First Gillard Polling
Today brings about the first robust polling of the Gillard era (there were a few not so robust jobbies floating around yesterday in the media using instapundit online panels – ignore them, they may well have had big sample sizes, but their sampling frame isn’t up to the job required for political polling.)
We have a Nielsen in Fairfax and a Galaxy in the News Ltd tabloids. The Nielsen ran from a sample of 933, giving us an MoE that maxes out around the 3.2% mark (demographic tables here), while the Galaxy was a sample of 800 giving us an MoE that maxes out around the 3.5% mark.
The first thing to mention about these polls is what they don’t measure. As a result of them being in the field on Thursday night in the case of Galaxy and both Thursday and Friday nights in the case of Nielsen – they wont be measuring a proper voting intention, but instead measuring an initial public reaction. We’ll have to wait for a week or so for the phrase “Prime Minister Gillard” to sink into the public brainspace before we can start to get a confident grip on the voting intention fallout.
As a result, don’t be surprised to see quite a large variation in polling results over the next polling cycle.
First up, the headline figures for the pollsters came in like this -I’ve also got higher resolution Nielsen data, so we’ll run the Nielsen results to a decimal place.
Earlier this month, Nielsen came out with a two party preferred of 53/47 to the Coalition – a few points undercooked for Labor based on our polling trends at the time and a result not replicated by any other pollster. So the size of the turnaround here in terms of initial public reaction is probably slightly overcooked for Labor. The last time Galaxy was in the field was mid-May – so using that as our baseline, Galaxy comes in with a two party preferred of 52/48 compared to May’s 50/50 result. If we look at the May result for Nielsen, it came in at 50/50 as well.
If we compare the primary votes of the parties and how they’ve changed since May, this is what we end up with:
Both pollsters show a drop in the Others vote and the Greens vote – although Nielsen shows a considerable Greens drop of 5 points. Also worth noting is how the Coalition vote didnt really move a jot on either pollster’s results. The initial public reaction to Gillard seems to be identical to the medium term public reaction to Abbott – it has encouraged skeptical partisans and ordinarily soft voters back into the fold.
You might remember recently we were looking at how Greens preference flows had been changing over time – with increasing numbers of Greens voters giving their preferences to the Coalition, dragging Greens to Labor preference flows down to as low as 68%. Today’s Nielsen poll has those flows bouncing back up substantially.
Meanwhile, the Better/Preferred PM beauty contest came in like this:
Galaxy also asked a few additional questions, the first if which was on whether the public agreed with removing the PM.
The next looked at responsibility sharing, where Galaxy asked:
Considering some of the controversial issues that have dogged Labor, including the delay of the ETS and the introduction of the Resources Super Profits Tax. On balance do you believe that Julia Gillard should share the responsibility with Kevin Rudd for the decisions made or do you believe that Kevin Rudd was mainly responsible for those decisions?
Finally, Galaxy asked:
In your opinion, which one of the two leaders, Julia Gilard or Tony Abbott, is best described by each of the following words or phrases?
I’ve highlighted the “trust” results which is still Abbott’s biggest weakness, especially now that Rudd – who was struggling himself in the trust stakes – has been replaced by someone who has a clean slate on that measure.
Please login below to comment, OR simply register here :
Thank you for registering, we have just sent you a confirmation email, which includes your new password to be entered below.
So will you include them in your pollytrend mix?
I will, but I might give this polling cycle a transitional weighting to reflect the uncertainty properly.
Thanks, Possum, a rigorous analysis, as usual.
I certainly agree that the first week, or so, of polling will reflect a ‘novelty’ component for PM Gillard, so we shouldn’t go too far overboard at this time, but the low point for ther ALP has been passed (and was passed, perhaps, before PM Rudd was jettisoned, but that is another calculation) and the rebuild of the primary vote to a steady set of numbers with a ’4′ in front, rather than a ’3′ is the necessary prerequisite for the ALP to be sure of retaining Government in the coming election.
Poss – do you find that those “uncommitted” numbers in the final graph tell a story?
And how do they overall compare to the last head-to-head between Abbott and Rudd? In other words, has the ascension of Julia changed Abbott’s numbers substantially and if so, for the better or worse?
They might Chinda – we’d need to see earlier versions.
I’ll go hunt for some.
I have a problem with understanding the changes to the Greens preference flows.
A number of Greens voters seem to have moved to be Labor voters. I would assume that these are voters who would have given their preferences to Labor, anyway.
Consequently, I would expect this to dilute the Labor orientated Green preferences i.e. less of the remaining Greens woill give there preferences to Labor. Hence the preference flow should have dropped rather than risen ??!!
Similarly, in the past, as the Green vote increased, at the expense of Labor, I expected the Green preference flow to labor to increase from the 2007 election result, rather than fall (as it did, to 68%).
Can anyone explain what is actually going on ?
Runim, what looked to be happening was that there wasnt a simple Labor-to-Greens change.
What was happening was that some Labor voters went to the Coalition and some went to the Greens, while some Coalition voters were also moving to the Greens.
So the headline numbers showed Labor down, Greens up and the Coalition being pretty stable.
But underneath you had larger amounts of Coalition voters ultimately ending up in the Greens column – diluting ALP preference flows.
Effectively, the Greens were getting a lump of swinging voters on top of their usual 8-10% vote they’ve been getting for 2 years, a lump of voters that then preferenced the least worst choice of major party with their two party preferred preferences.
So we ended up with the change in Greens preference flows over time starting to correlate with the major party primary votes.
runim, Possum’s preference flow analysis was interesting, but the trend that it plotted of a falling Green to ALP preference flow was based on only five sample points. The first was unusually large, the last unusually low, but both reasonable considering the MOE on the small sample. Take out those and the remaining three had a pretty much flat trend. This new polling would tip that trend upward a little, but again is within MOE (I think).
I don’t think you can extrapolate much from this at all. It’s mostly just noise.
My totally unrepresentative and unscientific water cooler polling (N=20) shows that:
100% of women (n= 12)polled thinks Rudd had a raw deal and may not even end up voting – these previous labor voters are appalled by the machine crushing a populary elected Prime Minister (and yes I know all about the westminster pretence at representaive government, the americans also do not vote for Obama either – they vote for electoral colleges and at primaries)
50% of males (n= 8 ) just shrug their shoulders and 50 % won’t be voting
Hmmm, sounds like someone is trying to hide the decline…
If Labor had left Rudd in and they lost- WE would have to live with PM Tony Abbott and his religion for 3 years min ? WE don’t deserve that! The money has gone to Julia , who gets it right on election day ? Money or Polls ?
On 8% of a sample of 930 (fewer if that number includes no opinions) it’s a sample of just over 60 Greens supporters. How much store can be put on whether such a sample came out at 65 or 80%
I think we need two or three new verses of “I’d love to have a beer…”
For the uncommitteds:
I’d love to have a beer with someone
I’d love to have a beer with some’
We drink in moderation
I dunno who with ‘cos I’m fairly dumb
We drink at the Town and Country
Where the atmosphere is great
I’d love to have a beer with someone
‘cos someone’s me mate
Graeme – there’s certainly some uncertainty there. But we would have expected the pref flow to the ALP to rise on the basis of the consistent pattern we’ve been seeing of late with those Greens prefs, so the number comes in as we would expect it to.
If the next poll shows the same pref flows from an ALP primary of 43 or more and a Greens of 10 or so, then the “its all random” argument would have pretty much run out of steam
The Green polling will come back to Gillard until she drops the mining tax to 22% then it’ll swing back to 14% for the Greens once the election is called, … my straw poll show an almost exact replication of 1GMD.
Thanks Mr Comitatus*. Yes, I meant rubberiness not randomness.
Wonderful Radiolab programme on the stochasticity of life, rebroadcast on the Science Show today. http://www.wnyc.org/shows/radiolab/episodes/2009/09/11
* Forgive me if that’s sexist. But since last Thursday, psephing may be the last important role not conquered by a woman.
Graeme @ 18 – Have you never heard of Enid Lakeman?
HARRYBELHARRY: Yes, we would have to suffer Tony Abbott’s religious cult for three years.
But it’s far worse than that, the Liberal Party is loaded with Oz Taliban members. Kevin Andrews Jesuit trained radical anti-abortionist, and anti-euthanasia; Chrissy Pyne, Catholic; Barnaby Joyce, National Party, Catholic anti-abortionist; David Clarke aka Ratzinger’s Right; Peter Ryan, anti-abortion National Party; Cory Bernardi, born Catholic, since turned to be Anglican; Steve Fielding, Hill Song? Moron? Seventh Day Adventists? So forth and so on.
It’s easy to say the numbers, compared with the overall numbers are small. But note they are in positions of potential power.
It used to be Labor who had the worst examples of Catholicism-and they indeed have some monumental stinkers-Steve Conroy comes to mind.
It is the nature of any far right-wing fundamentalist-Islam, Anglican, Methodism, any religion-to want to convert the rest of the electorate to their way of thinking.
It is to be hoped our fragile constitution can withstand the tidal-wave of religion coming with Tony Abbott, and his Catholic cohorts.
Thank the relevant god that Julia Gillard is now the Prime Minister. The religious right had a massive cardiac-arrest because she, at her swearing-in, refused to swear on a stack of six-ten thousand year-old fairy tales. She is an atheist. Olé!
The next salvo from the religious right will come in the form of “She’s no fit person to become PM because she hasn’t had any children!” after that they will “Profess to feeling shame because she isn’t married to her partner!”
I’d give it about a week before the first salvo is lobbed onto the deck.
Hi possum.
Am a newbie here and don’t have much knowledge of how to read these polls.
But on face of it, surely Neilsen doesn’t have good news for Labor?
If you take out the clearly OTT Victorian thrill (Lab primary 59%, 2pp 67%) aren’t you are left with NSW, QLD, SA,NT and WA … where ..
Primary: Labor 41.5, Coalition 48.25 (!)
2PP: Labor 48, Coalition 52 (!)
Galaxy poll seems more realistic with a 2PP Labor 52, Coalition 48. But given the Vic thrill, can’t it be assumed that it must be skewed by Vic too and the other states are still running hard for the Coalition despite Julia’s elevation?
Is there some explanation to do with analysing poll data that justifies the headline nmbers out there?
Welcome aboard Bar Bar.
Rule number 2 of polling: Sample size matters (I should write these rules down somewhere)
The larger the sample size, the smaller the margin of error – so the larger the number of people polled, the closer we expect the true value of public opinion to be accurately represented by the poll.
When we get to the geographical cross-tabs – like the state level polling results that make up the national results, that rule still applies.
So with a national poll of 933 people – as Nielsen was – the margin of error is up to 3.2% at the national level. When we drill down to the state level, the margin error on those results are 5.5% for NSW, 6.2% for Vic, 7% for Qld, 9.8% for WA and 10.6% for SA.
So the state numbers are very uncertain.
During periods of polling calm where Galaxy, Nielsen, Newspoll, Morgan and Essential Report all get results which suggest there has been absolutely no movement in either the primary or two party preferred results over a two week period – if you drill down into the state level data of those polls, you’ll find wild variation week to week, where SA might be 57/43 one week and 50/50 the next, NSW might be 54/46 one week and only 51/49 the next.
Yet, ultimately, all of that state-by-state, week-by-week variation comes out in the wash to give the same national primary and two party preferred results.
That simply comes about because of the much smaller sample sizes attached to the state level breakdowns, where the smaller sample sizes give us larger random movement around the true value of public opinon.
Now there’s no doubt that Victorian number is overcooked for Labor – but other state results like SA (for instance) are probably undercooked.
So the national level is what we need to pay attention to – and once we get enough of those polls (say, 3 months worth), then we can break the aggregated results of those 3 months worth of polls down into smaller cohorts like State level, and age group etc and be much more certain of their accuracy.
So, if an election is held in August or September, we should have enough opinion poll data available to decide who won and where within about a month or two afterwards, Poss?
Cheers
Rod
Sample size. Thanks for that. But if it takes 3 months of polling to get more certainty of accuracy then I ask myself why these small sample individual polls are not issued without heavy qualifications and in 3 months time wouldn’t they ony make sense if the same number of people were polled each time? ie comparing like with like?
And also, one state is wildly out of kilter. The others, including NSW, taken together, appear to be in the same range.
Sigh, I suppose there is some mathematical statistical formula applied. But these quicky polls appear very iffy to this poster?
Ignore any backlash over the method of the changeover. If voters could forgive Malcolm Fraser for an orchestrated coup in 75 and vote for him shortly afterwards, they sure as eggs aren’t going to “maintain the rage” and install a raving right religious clown like Abbott.
I think that the election should be held as soon as possible, while the Libs have no policies. Maybe bed down a compromise RSPT, get some kind of machinery in place to revive the ETS, wait for a few polls from which a trend can be drawn, but otherwise catch the Libs with their cassocks down (or would that be up?).
Heavy use should be made of that disgusting Abbott presser after the change of leadership. A sneering, nasty piece of work that was. Linking Abbott to a UK style horror budget as well as the return of WorkChoices would also be de rigeur.
http://www.roymorgan.com/news/polls/2010/4517/
And here is the tragedy of their panicked stupidity.
I despise Gillard and I am glad I have not voted for the ALP since 1983.
While I, on the other hand, couldn’t stand Rudd, and would not have given the ALP a preference with him as leader. His rampant xtianity, and his habit of presumptuously “talking for all Australians”, the neglect of the “unworthy poor”, and many other aspects were really irksome. Not the least his close resemblance to the previous incumbent on many issues. Now I feel comfortable in giving the ALP a preference ahead of the Libs on a formal ballot paper.
@27 So you’d rather have a party that screws the whole country than screws its leader? Excellent logic there
@28, just out of curiosity, do you think the personality of the leader actually has much effect on real policy decisions? I certainly have some criticisms of Rudd, but I remain to be convinced that Julia will undo the policies aimed to please the wowsers.
Maybe not their personality, but their religious and other beliefs do.
Well, there is at least SOME chance that that may be the case, rather than none under Rudd, and even worse if Abbott ever got the job.
There is always the possibility that Gillard could soften any policies proferred by the right. No such possibility existed with Rudd.
IMDG – I would suggest that your water cooler conversations must have been with LNP voters as judging by the above figues there is much more concern about the change by LNP voters than by Labor. My experience has been the exact opposite and the Greens that I know were definitely favourable to the change. People may not agree with something like that but it was handled very well by both Rudd and Gillard. The swap whilst obviously emotional was not damaging in the way that a lot of build up name calling can be.
I agree with Venise, the good thing about the Gillard appointment is that it’s one in the eye for the religious establishment. I see she has already watered down Rudd’s Big Australia (a cherished christo-capitalist cause). Of course her smaller Australia will probably still be too big…
The Galaxy question on whether the decision to change leaders was a good one seems very poorly written (“they panicked and made a poor decision”). Why not just ask if respondents thought that this was a “good decision”?
Likewise, labeling several policies as “controversial issues which have dogged Labor” in the introduction to the question on whether Gillard shares responsibility for these issues is unnecessary and will probably influence the results – why not say “controversial issues”?
Poor Ruddy – he learned a little lesson in politics – did we? – poll at: http://wp.me/pXIwk-1z
Why does Nielsen show a big change in the Green vote but Galaxy only shows a small one?
Raw Prawn – no particular reason, polls often show variance.
It all comes out in the wash over an arbitrary period of time. Expect to see bit of variation in the polls over the next few weeks as the public adapts to a changed PM.
If having a first female PM is Labor’s platform for another term of Government, then surely the Liberal Party must level the playing field and have Julie Bishop as Leader and put Abbot 2IC.
What we must understand is that this Labor Government is devoid of any real sustainable policy and has mercilessly destroyed the Australian economy with flawed policies and taxation reform.
Certainly, the Australian people are discerning enough to know that Gillard has been one of the prime architects of Labor policy failure.
I say to the Libs – Its time to give the lady with poise and class a go – JULIE BISHOP for PRIME MINISTER
hahaha Ball – you must be joking!
Your second and third paras though show that you are a rusted on blue meanie.
Please enlighten us as to what policies “Action Man” has other than “block block block”, and “sink the refo boats”?
I must admit Abbott sounds as if he’s panicking badly to the change of PMs. His responses to Gillard’s statements sound even more childish than they usually do (if that’s possible).
Where are you teaser of the dangles? I am also interested in your reply to DavidR’s question The man of God this morning I note, took credit for bringing up the smaller pop growth first, but when questioned further on Insiders he didn’t have any specifics, just that he thought of it first and, wait for it, Opposed Rudds model. Mind you this was not a scripted interview so!!!!! ……………remember it came from his own mouth.
Tomorrow I’ve got new sidebar demographics and election simulations from the last quarters worth of polling data – it’s rather interesting
calypto hey he’s staked his political career on those hugely witty “Kevin the Lemon” commercials. You’d panic too.
@ Possum,
Is that just Rudd figures?
@35. twaddle
The main trend I’m looking for is how Queensland goes. The swing at the last election was so large in some of the unexpected Labor wins there that, with the loss of a Queensland PM, the reverse swing could be equally as devastating if the Coalition plays that card. The lack of movement in the Coalition primary is also notable.
Also, the Right wing of Labor making all the running in this move is no surprise; not because they are so machiavellian but because if the Left tried, being used to losing, they’d be laughed at, and the Right, with their win at all costs motivation, were the only faction left standing. So essentially it could only be the Right who would try.
John,
Yep – just the figures that Rudd had garnered over the last 3 months
Pedant@19. Yes, Lakeman, a Brit woman and liberal who championed STV. Didn’t realise she was a pseph in the quant sense.
I meant I can’t think of such a pseph in Australia: plenty of local women have shone as scholars of parties, electoral rules and psychology (eg Marg Cribb, Lisa Hill, Joan Rydon)
“Despise” is a very strong word, shepherdmarilyn, and it is hard to fathom why you might develop such strong feelings about someone who has been the Prime Minister for a matter of moments unless you have some sort of highly personal beef with her?
I guess if you haven’t voted ALP since the “drover’s dog” election in 1983 you aren’t likely to this time whoever happens to be leader? Heck, Gillard was still at Uni back in those days!
So who have you been voting for in the years since Hawke’s victory over Fraser? The Dems and then the Greens, or the Libs and One Nation?
Good to see Julia dumped the idiotic “Big Australia” policy – one of Rudd’s biggest blunders.
Tony Abbott was being less than truthful in claiming that he had the idea first – he was actually in favour of a big population for Australia
3 Trackbacks
Please login below to comment, OR simply register here :
Thank you for registering, we have just sent you a confirmation email, which includes your new password to be entered below.