While the campaign has been fiercely fought, there hasn’t seemed to be much for myself to add in terms of blog posts on the front page.
The conversation has been continuing fiercely on the previous election thread and on each seat’s discussions. Almost 2500 comments have been posted throughout August.
Please use this thread to continue discussing the election campaign over this last weekend of the campaign.
Thanks Ben for a great forum.
I am the Independent candidate for Menzies, VIC. If anyone has a question for me I’ll be happy to answer. It’s been a baptism by fire, I have been learning about: the constitution, clarifying my message, local differences, preference strategies (ALP put Lib 3rd here after GRN), money & marginal seats, letter dropping, coverage & density, brainstorming techniques, minority & special interest groups, email campaigning, sore feet, …
Well done Ramon, for putting your hand up. It is an experience that few are prepared to even try. Plus you get to see it all from the other side. The preference system is open to abuse and manipulation, and I have not studied the alternatives enough to suggest a better system. Any suggestions from those more knowledgeable than me?
I think yesterday’s clanger over the Liberal costings has probably cost Labor a few seats. For all the noise in this election, That one WILL hurt.
DB – it’ll only hurt them if they don’t give the right answer when asked about it. And sadly, at the moment they’re not giving the right answer.
When asked why the Treasury officials are putting out those statements, Labor should be emphasising that they’re simply clarifying the situation to make sure that there’s no doubt about what is actually happening.
When asked how they can claim to be identifying a $10 billion “black hole” (I hate that phrase in the context of finance, no matter which side uses it) based on policies that the Coalition claims aren’t representative of the real policies, Labor should be saying this: “The Coalition is doing one of two things – either they’re lying about the costings of their policies, or they’re leaving out key details of their policies that significantly impact the costings; in either case, they’re lying – directly or by omission. But we would be happy to be proven wrong, by the Coalition releasing their costings. We put in the request for costings based on their announced policies, which are the only policies that they have explicitly put forward to the people.”
That they aren’t doing this saddens me.
I did not think the campaigns of the parties was anything special. Lots of party PR types organising PR events attended by only by the media mostly. While the rest with kissing babies, collect fruit from the market, standing in a shelved warehouse or wearing funny helmets and yellow jackets is all so fake and meaningless.
At least the pollies have stopped staring up at a car’s brake drum on a hoist. That was a real favourite but with the car industry in decline perhaps they realise that there are few votes in it now, if they ever were.
I hate presidential style campaigns. The recent vote on Syria in the House of Commons proves the Westminster system is best rather that a US President making a decision without a vote in the Congress
Glen, I’ve looked at all the front pages of all the national papers today and this is going to hurt as it goes to the Government’s credibility. Also had a look a Tim Colebatch in Fairfax and for him to say it was their Ralph Willis moment tells me this one is going to hurt. I find Tim to be pretty much anti-Abbott but this analysis is politically correct.
People don’t investigate to the level you are suggesting. They look at the headlines and make a decision from that. It is only the political nerds that look into the detail.
It’s worse than you say though Glen because the Government needs to backtrack substantially for any credibility in this matter (dare I say it is a bit of a Mark Latham moment in a different way). The PM said the Coalition were “committing fraud on Australians”. That Glen is NOT going to go down well without absolute proof and justification on their part (not the Coalition’s) and Treasury have refuted these assertions of the ALP as to costings. That is what people will remember this late in the campaign.
Effectively the ALP’s comments yesterday have probably allowed the Opposition to not release their full costings until very late next week and it would appear the Opposition are now in a strong position at this point in the campaign.
DB – what I’m saying is that it’s hurting because Labor isn’t answering the questions correctly. There’s no need for people to do research for Labor or the issue, just Labor actually emphasising the right point about their Treasury Costings of the Coalition’s announced policies.
Note the use of the word “announced” – it’s a key one that Labor isn’t emphasising well enough. People (which mostly means journalists, who then distribute the key info) can’t judge the Coalition on policies and policy details that haven’t been announced, and if those details are so influential as to make a difference of $10 billion, surely that’s just as much a problem as if the Coalition’s costings are out by $10 billion.
There are a total of three possibilities when you break it all down – either the Treasury figures are just wrong (which would require explanation as to why they disagree – the Coalition should be pointing to the flaws in it), Treasury figures are based on announced policy rather than internal policy (which would require there to be key details that the Coalition have not released), or Treasury figures are accurate (so there’s a big gap in costings). Or a combination thereof, of course. But Labor just keeps parroting their pre-decided lines, and that’s why I’m so annoyed at it – that they’re focusing on attack rather than scrutiny. Note that the Coalition aren’t doing any better.
Peter Brent seems to back up my claims in Mumble today, but said differently. An extract below.
“This week, even without the bureaucrats’ unusual intervention, the government’s announcement of a $10 billion black hole in the opposition’s costings was confusing politics. For weeks and months they’ve been ramming the $70 billion figure down the throats of Mr and Mrs Semi-Engaged-in-the-Suburbs. Now it’s just a fraction of the amount?
And what’s $10 billion in the scheme of things—the hundreds of billions of accumulated debt?
…..Thursday’s letter from the Secretaries of Treasury and Finance departments,……..Put it this way: would the mandarins have released that document if they if they expected the Labor government to be returned?
Australians are used to their politicians being loose with the truth, but this is so damaging because it sprinkles salt onto that festering area of vulnerability, the economy.
All for just $10 billion”.
I’ll repeat this bit from Peter Brent:
“but this is so damaging because it sprinkles salt onto that festering area of vulnerability, the economy”.
Remember, this is Peter Brent, not me. But Peter Brent and I are in total agreement, again.
Now me. “It’s the economy stupid”. It’s a simple line but it’s what decides elections. And it’s going to decide this one. I reckon I know which way it’s going to decide it. Yesterday’s decision by the Government to attack the Opposition only to have it backfire badly as a result of Treasury’s statement is a strong turning point in this election. It is first real monumental stuff up on either side. Given the comment and accusation by the PM, I find it very hard to believe the ALP can win from here, and any slight improvement to the ALPs primary vote will most likely be an abberation.
Is this Godwin Grech all over again? Only time will tell.
DB, in response to your question ” is this Godwon Grech all over again?”, my intial thought when I saw the news last night that is was deja vue GG. Rudd had completely over extended like Turnbull and when it turned out to be “untrue” then there is no where to go. A leader can;t ask for the “resignation of a PM” or call the opposition as “perpetrating a fraud on Australia” with out absolute proof. It goes past the normal rough and tumble of politics. I think it reeked of desperation and Rudd is now trying to save the carpet having lost the furniture.
Seperately, this election seems to be the least engaging for as long as I can remember. It was news item 3 on both 7pm ABC and Nine 6pm news last night. The “debates” have been on the secondary digital channels. Rudd hasn’t performed well but Abbott hasn’t done anything to turn voters off (despite any misgivings they may have). Accordingly, the majority of the swinging voters will vote liberal (or preference liberal before labor) and the coalition will win.
Glen,
I think the problem is Labor’s already shot their bolt on the costings issue by Rudd going so hard so soon. It’s a bit late for them to push a balanced, reasonable case over the issue now.
As an aside, interesting article on how the AEC’s recent redistribution for Victoria has gone horribly wrong. All federal electorates in Victoria should be approaching the same average number of voters within the next 12 months (within 3.5% tolerance), but 30% of Victoria’s seats have now no hope of meeting this.
http://www.aec.gov.au/Enrolling_to_vote/Enrolment_stats/elector_count/2010/index.htm
The variance is already quite marked with McEwen aready 17% over quota (only three years after the redistribution) and is already the nations largest electorate with 111,000 voters.
So it will be interesting see how this effects the result on election day and whether it effects the margins we’ve been assuming in some of the high growth seats in Victoria.
Evidence of Liberal candidates refusing to turn up to public forums continues to accumulate – couple of recent examples in the ACT and surrounding region since the initial reports from Sydney – this is treating the voters with contempt.
An ALP Senate candidate was not available in Victoria for the forum next Monday (02 Sep 13) at St Kilda Town Hall – 6.30pm for 7pm. All welcome. Google “locals into Victoria’s environment” for details. 7 Senate candidates (all different parties) are attending including a Liberal and a Greens.
Doug – no, it’s a smart move mate.
Morgan is already out with a new multi-mode, and they show the Coalition up 1.5 points on half a week ago on both measures of TPP (to 53% on respondent allocations and 54% on 2010 preference flows).
GhostWhoVotes gives us some data from Newspoll in Banks, Greenway, Lindsay, Parramatta, and Reid. The Coalition is on 57/43, and 71% of Coalition supporters are certain they will vote that way on the day, against 53% for Labor. Uncommitted were at 4%, suggesting that Glen’s theory about late deciders going to Labor, if correct, won’t help in these seats.
no one seems to have mentioned the latest western Sydney poll showing a 9% swing against
Labor was a robopoll that is no person asked the questions. my reponse to such a poll would
be to hang up…..I suspect a high non response rate….. would not this make the potential
margin of error higher…………the polling companies get a quick result this way…..
but it means very little
MDMC – Oh, I agree, Labor have been playing the entire campaign poorly, and their costings angle was a bad choice of attack – because it basically encourages the Coalition to release the costings earlier to prove Labor wrong, whereas Labor could have done a last-minute “Why has he been keeping the costings hidden all this time?” angle in the last week to hit the Coalition hard.
If I were running Labor’s campaign, I’d have done it very differently. I’d have had the continuing MPs and candidates remain completely positive and focused on Labor policy, and organised for the retiring MPs to do all of the attacking – Gillard, Emerson, Smith, Crean, etc. I’d also try to get Bob Hawke to be the face of the attack. Played right, it would be incredibly effective, in my opinion.
I’d have had Rudd put the election off until October, rather than rushing to it, and have parliament return, so that they could launch a heap of legislation that would assert incumbency while embarrassing the Coalition (by forcing votes on issues that split the Coalition, for instance). I’d have also used the time to ensure that every division has a Labor candidate up and running, and where candidates were poorly selected primarily because it was an unwinnable seat, I’d have pushed to have the pre-selection restarted to ensure that the best candidates were available.
I’d have had policies focused on health, education, etc, and put those at the forefront, with a reminder that debt isn’t the bugaboo that some make it out to be – it’s something to address, but should not be placed ahead of making the country a better place.
The problem is that Labor seems to be trying to be the right-wing party, and they don’t quite have the stomach to go all the way. When people have a choice between imitation right-wing and genuine right-wing, they’ll go genuine every time. They’re losing votes to the left and failing to pick up votes to the right. They need to re-evaluate their approach.
Glen, right on many points.
Ok, out on a limb here, and maybe I drank too much of the Cool-Aid, and I know most people will laugh, but here it goes. KAT/PUP to win 6 to 15 LH seats. Now, if I got it wrong, who cares. But, if I get it right, I get bragging rights.
What do I base this on? Some exit polls at the pre-polling, and feedback from a few people.
Hey, shoot me down. I am big enough to take it.
Cheers
RichR – the Morgan poll has Coalition primary numbers down 1%, and Labor primary numbers down 4%, with Greens gaining 1% and “Others” gaining 4%. That’s a pretty sharp swing in numbers – in none of their other polling during the election period has there been such a sharp swing for any party, let alone two of them. There have to be outliers, and I think this might be one.
I looked into it a bit more, and this “multi-mode” only has internet and phone polling. This is as opposed to the usual trio of SMS, Internet, and Face-to-Face. So it seems that this poll isn’t consistent with their usual process, and this would explain the difference in the numbers. I kind of wish that Morgan would provide break-downs of their multi-mode polls based on which mode was used. I’d be interested to see if face-to-face is biasing in one direction, etc.
Also, the explicit “undecided” number isn’t the only valid one in my conjecture. There’s also the “non-response” number. People who are undecided are less likely to respond to an opinion poll for obvious reasons.
Rockman – not going to “shoot down” your prediction, but provide a little info that might make you tweak your prediction.
It’s very likely that voters for substantial minor parties (Greens, KAP, and PUP, mostly) are more likely to pre-poll vote. In 2010, the Greens got 15% of the pre-poll vote, but only 11.5% of the ordinary vote. I’d expect a similar effect occurring with the KAP and PUP.
Glen,
I think it’s pretty well excepted that F2F favours Labor, which I think might be why Morgan changed to multi-mode in the first place.
I’d expect a phone/internet to be more in line with the actual trend than a F2F/phone, which I’d expect to be a couple of points pro-Labor….looks like that’s what has happened.
Glen, I provided the Morgan poll for general information, not as a point either way in the earlier debate. I take your point about the change in methodology making comparison dubious. I didn’t know that was the case at the time. As for the non-response rate, that’s true as far as it goes, but in this case that was 1% for a total of 5%, meaning it still wouldn’t be enough to overturn the TPP figures (assuming for argument’s sake they are accurate) in the unlikely event all of that went to Labor (which I don’t think even you are arguing will happen).
On the mismanagement of the campaign, I think going to Rudd was a mistake. It is unfortunate that Gillard was never able to communicate properly with the public, and I think her greatest fundamental problem was that she thought she needed to work as hard as she could to keep the last parliament together for a full term and accomplish as much as she could. That of course led to deals she didn’t have to make, including one that made her violate a campaign promise and one that she couldn’t keep up her end for. She should have forced the Coalition to vote against popular things and tried to engineer circumstances where she could go to the people on the basis that the important, popular initiatives were being obstructed and only a Labor majority could deliver. Her impulse was, I think, noble in its way, but politically unsound.
Given the reality of her unpopularity, Labor should have turned to someone new early this year. Unfortunately, the Rudd-factor sucked up all the oxygen, and there was never really any chance with him in the picture that anyone else could get up. In the end, Labor rewarded bad behaviour to save the furniture. It’s impossible to say whether Gillard could have gotten Labor’s numbers back up during an actual campaign (it’s been done before), but only someone new, someone untainted by the backbiting of the last few years, could have actually put Labor in a winning position. By that I mean a position where Labor could win on its own merits rather than hoping for a massive stumble by Abbott and the Coalition.
Incidentally, and this goes to Observer, it is this antipathy toward Rudd (and affection for Gillard) which led me to say (as I said at the time) that the disability scheme wasn’t a strong accomplishment for him to run on. It wasn’t his accomplishment (and was supported by Abbott anyway).
The prevailing view (on this site and elsewhere) seems to be that a Coalition landslide is inevitable. I just don’t see that happening. I think that people have come to the view that a landslide is on the cards by reference to polling. However judging the validity of polling by reference to how well pollsters predicted the 2010 result is flawed because of the massive shift to mobiles and away from landlines since 2010. There is also a disconnect in my mind between my lived experience of Qld parochialism and the national polling results in Qld on the one hand, and the marginal seat polling in Qld. I think the marginal seat polling in Qld is unreliable (for example, I flatly refuse to put any weight on polling produced by the firm that suggested KRudd is going to lose his seat – what rubbish).
To be clear, I think the Coalition will win, but I think that some of the predictions about ALP seats to fall are fanciful. I see Labor losing around 12 seats in a best case scenario for them (treating Lyne; New England; Dobell and Fisher as notional ALP seats for this purpose). However, I also think that the ALP will pick up probably about 4 seats in Qld (Brisbane; Flynn, Forde and one other). Something else that hasn’t attracted much attention so far in coverage is the situation with regional Qld seats (such as Herbert; Hinkler or Leichhardt). I think there is a good chance these seats will be lost by the LNP to Katter (or possibly if the ALP is lucky, won by the ALP on Katter preferences).
On this analysis, if absolutely everything went right for Labor, the Liberals would have a narrow majority with around 77 seats to Labor’s approx 68 seats with 5 KAP/independent/Green MPs. I think there will be a shift toward Katter and independents so that there will be at least 5 independents elected and I think Bandt and Wilkie will retain their seats.
On my worst case scenario for Labor, they would lose around 17 seats, and the Libs would also keep all their Qld seats, leaving Labor with only around 58 seats. So I guess that leaves me predicting that the ALP will be left with the mid point of 63 seats. However, because I don’t put any weight on the marginal seat polling, I lean towards the view that ALP seat numbers will not fall below 61. On this basis, I think the betting odds available for some ALP marginals are very enticing (for me at least), as are the nearly even money odds available for the ALP losing by less than 30 seats.
I think the below ALP or notional ALP seats are gone:
Banks
Lindsay
Corangamite
Deakin
La Trobe
Braddon
Bass
Robertson
New England
Lyne
Fisher
At least one more will fall from Dobell, Reid and Parramatta. I don’t know much about SA, NT or WA so I can’t really comment but overall I don’t expect much worse for the ALP than a net loss of one seat across those states.
Seats that are talked about as ALP losses that I don’t believe will happen are:
Eden Monaro
Kingsford Smith
McMahon
Page
Watson
Werriwa
Richmond
Not only is 07 Sep 13 “Endangered Species Day” according to my dairy but it is also St Crispins Day (Wil Shakespeares – Henry V play/film). Remember that great speech to the smaller English Army before the defeated the French Army. Rudd is not Henry V though.
MDMC – the point is that it’s not following the same polling technique, and thus any perceived “swing” cannot be considered accurate. Each technique will have a bias, but one can still identify trends from them, generally. Whether the numbers themselves are more accurate is hard to say, but the apparent “swing” of 1.5% to the coalition isn’t a real swing – it’s a result of a difference in polling technique.
RichR – I think Labor should have switched to Rudd earlier. He was more popular with the public for a reason. The problem is, by waiting until the 11th hour to switch, they were focused on the short-term bump from the switch rather that the long term potential that could have come from establishing incumbency. His actions immediately after being returned to the leadership were laudable – primarily, the NSW Labor intervention, the changes to the Labor rules regarding leadership, etc – but they weren’t given enough time to really establish themselves, and thus the value arising from them was limited.
Sandbelter: McEwen isn’t the nation’s largest electorate. Canberra (127k electors) and Fraser (138k electors) are significantly larger (and more than twice the size of the smallest, Solomon).
Correct,
I overlooked the the two ACT seats as it reflected a political decision to restrict the number of seats in the ACT. But the variance between the seats is tolerable given a redistribution is imminent for the ACT
The point I was trying to make is the variance in Victoria reflected the botched job the AEC did in the recent redistibution. Based on current growth rates McEwen will pass Canberra by the next election (if it goes full term) anyway, Lalor may follow shortly thereafter. Meanwhile Aston, Deakin, Bruce and Chisholm will will see falling enrollment in the 90-95K range.
Sandbelter
See my post to you in/on Melbourne Ports. Perhaps there will be very few electorates with falling numbers due to the new automatic compulsory enrolment. The problem with McEwen was caused by the AEC failing to proceed with the new seat of Burke (& abolishing Murray). A bit like doing a renovation without any demolition. The design is instantly flawed.
Well I am predicting at this moment the Coalition to pick up 10 seats in NSW and 4 in VIC and at least 2 in Tasmania. The rest will be about breakeven, although QLD is hard to get any confidence about (it could well go a lot to the Coalition and I think this is a bigger chance than Labor picking up a bag full). I expect the Coalition will win between 85 and 90 seats on around 53-53.5% of the vote.
I am just hoping the Coalition get the chance to deal with either the Greens, Labor, or Others in the Senate.
MCBAL, I would have Reid before Banks. Banks has never been won by the Libs, yet Reid (under the old seat) was a bit of a swinger in the 90s. I do think Banks will go too though.
I’m not going into specific seat by seat analysis, but I think many Labor supporters are way underplaying how bad it’s going to be in NSW. Obeid, Thomson, O’Farrell’s popularity (and don’t worry, he is popular and still pulling 60/40 State polls) are all going to hurt Labor in NSW more than I think many people believe. Those Newspoll’s in the marginal areas are not off the mark from what I see and they are occurring in areas where O’Farrell particularly has laid his transport footprint.
The ALP will be lucky to get 45% of the 2PP vote in NSW in my opinion and they overreached in seat terms massively in the last election on around 49% of the 2PP vote in NSW. I don’t think that can be repeated personally, which is what you are suggesting is going to happen.
As to the Senate in NSW, I do see a 4th right Senator as possible with the Libs/Nats getting 3.
DB, I agree with your assessment on Labor losses in NSW, VIC and TAS.
I am prepared to make a likely prediction in Queensland based on betting odds and polling in recent weeks showing Labor not winning any LNP seats and not to retain the seat of Moreton.
My gut feeling is that Labor is also in trouble in Blair, Capricornia, Lilley and Petrie. A poll published in the “Courier Mail” today shows the 2PP vote in Blair is 50:50. Unless there is an improvement to their vote in the final week they will lose some or all of those seats. Labor on track to receive a 2PP vote similar to the one received at the 1996 election.
Lurch – all of those polls are untrustworthy. In Griffith, one poll last week said Rudd would lose his own seat (by either 2 or 4% margin for the Coalition candidate). Another one more recently said he’s ahead by 7.2%. A swing of more than 9% in a few days? I doubt it. Robopolling hasn’t proven itself to be accurate, yet, and numbers haven’t been consistent in any way. You have robopolls of marginal seats giving swings against Labor of 6-14%, while national polls have Labor at more like 48%… the only way for this to make sense is if non-marginal seats (on both sides) are actually swinging to Labor even as margin ones swing strongly against.
Glen, I don’t really take too much notice of any one seat or Robopoll (take for instance McMahon today). But when you get poll after poll, one after another, saying much the same thing, then I think there is a swing on.
I think Lurch’s arguments are quite valid as are yours when it comes to individual seat polls.
As for the national polls, I don’t think they have Labor at 48%. Essential has 50/50, but Galaxy is 53/47, Newspoll is 53/47, Nielson is 53/47 and Morgan is 54/46.
DB, if the robopolls are mostly getting results inconsistent with proven polling techniques, which these certainly are, then it leads to doubt about their veracity. We know neither what the typical bias is within the method, nor what impact variations in the questions and options have (when a poll has an operator, there’s ability for clarification).
And I said “like 48%” for a reason. I went for something around the mid-range of the range that we see. Also, the Morgan number isn’t 54, it’s 53, and that’s from an oddball poll set that didn’t follow their usual modes.
It’s also worth pointing out (again) that even some polling with huge sample sizes a couple of days before the 2010 election were out by more than 2%.
Glen, that is rubbish. Except for some JWS polls, the remainder are well within the margin of error. Even Kevin Bonham has accepted that. And of course any poll can be out. That is margin of error. It could well be margin of error on the Coalition side or the Labor side.
And Morgan was 54% based on preferences from the last election which have been pretty consistent for the last 3, and given most preferences from the 2 major other parties are mainly flowing to the Coalition (KAP and PUP), then it could well be that this preference distribution is being understated. (By the way, that was with the online polling. I’d suggest that phone poll was actually about 57/43, but they didn’t want to position it that way).
The polls are what they are. Why argue that they are just plain wrong unless you have better numbers?
Whoops I think I was the first to release that Galaxy poll……
The last one prior was 52/48.
Everything except Essential is within 1% of each other: Morgan, ReachTel, Galaxy, Nielsen, Newspoll.
Seems pretty clear to me what the odd one out is.
The Essential Polls have tended to favour Labor though. The fact that the ALP is polling 50:50 on their most favourable poll suggests that they are in trouble.
Hawkeye, the current Essential status is weird even for Essential.
Glen, the Morgan poll is not “really” 53/47. Morgan is the only poll that reports a respondent-allocated preference as its headline number. The reason is that everyone else realises previous-election flow is more accurate, regardless of which party the differential favours going into any given election. Morgan is an eccentric operation, which can be seen in the arbitrary shifts in methodolgy (they’ve used three different survey methods during the campaign, two of them calling themselves “multi-mode”) and their often silly analysis of their own polling data. So claiming Morgan’s weird preference for how to distribute preferences gives you a “real” number, as though it is comparable to anything else happening in the world, is wrong. You can argue, as you have and with little reasoning, that something about this election is so different that we should expect respondent-allocation to be more accurate, but that doesn’t make it any more “real”.
DB, you guess may be right about it being 57/43 on phone responses. After all, when Morgan did a phone poll earlier this month, that is exactly what they came up with. Given that all the polls except ReachTEL were bunched around 51/49 to 52/48 at the time, there is little reason for confidence in phone results from Morgan. I suspect that that oddball result (for which they never released a 2010-flow TPP and have yet to list with their other 2010–2013 polling) is the reason this last poll had an internet component, but no F2F or text. In other words, it was supposed to be a phone poll, but they hedged their bets by diluting it with internet polling. It was, after all, released a few days after the last real multi-mode, which is a time-table more in line with Morgan’s phone polls, which tend to be released between their weekly polls.
DB, the robopolls have been reporting swings of the order of 10% in marginal seats where internal polling is saying it’s very close. They keep putting van Manen well ahead in Forde, for instance.
I find it hilarious that you refer to “even Kevin Bonham”, as though namedropping means anything to me. I’m a mathematician, meaning that my field of expertise is numbers and logic. Appeal to authority means nothing to me.
I’m not arguing that they’re wrong, so much as that they’re incomplete. I’ve identified a number of rational factors that explain behaviours seen in the polling trends that otherwise remain unexplained, such as sudden changes in polling numbers at a time when nothing of significance happened in the campaign or otherwise. I’m not asserting that the numbers are actually better for the ALP, just that we don’t know how accurate they are. Based on 2010 polling accuracy, the amount of confidence we can have is actually minimal – I mean, if a poll involving nearly 30,000 respondents can be off by 2%, if polls of marginal seats predict swings of the order of three times the magnitude actually seen in the election just a couple of days later, you have to wonder what’s going on.
I proposed an explanation for part of the phenomenon a week or so ago, and you said you were thinking about it. You have not yet provided any counterargument to it, yet you’ve since spent all of your time acting as though the polling is accurate… even as the internal Liberal polling you have access to disagrees with it in various places. It seems as though the only time you ever actually say that there’s any issue with a poll (outside of references to internal numbers) is when they say Labor is improving its position. Simply put, at this point I have to think that your attitudes and claims are based on partisanship rather than rationality.
Both leaders have been speaking recently about how polls typically shift in the last week, usually to rebalance towards 50/50. This is consistent with my hypothesis that one party tends to have their support solidify quicker than the other. In many cases, it’s no different from the past-voting effect, where people tend to claim that they voted for the winner, even if they didn’t.
Hawkeye – Now Essential is the one that leans to Labor inherently? I thought it was Morgan. In mid-July, Essential had Labor at 48% 2PP when Nielsen and Newspoll had them at 50%. For most of the last three years, Essential has been the worst for Labor of the three. It is disingenuous, to say the least, to claim that Essential disagrees with the others because it leans towards Labor in some way.
MDMC – you clearly haven’t ever studied statistics.
RichR – I’ve provided a HEAP of reasoning for my suspicions. Indeed, I’m the only person who has been able, it seems, to provide any explanation for a number of strange poll results that have happened during the campaign.
And past preference flows are only informative when you expect current preferences to mirror previous flows – which would only be expected if HTV cards remained consistent (both between elections and between electorates) and people who vote for a minor party always hold the same view of the majors at every election. This election is unusual for a number of reasons – a highly unpopular Liberal leader at his second election as leader is up against a popular Labor leader leading a Labor party that is highly unpopular at the moment. This is reflected, for instance, in the preferred PM measures, which have typically kept Mr Rudd ahead of Mr Abbott.
Then you have the big complications of KAP and PUP, parties expected to draw significant numbers but never actually polled properly (they almost always get lumped in with “others” – which will likely result in underestimation of their support) and whose preference flows can’t truly be estimated at this point in time. In the seats where KAP is expected to poll strongest, they’re preferencing Labor on their HTV. And these are seats where a lot of people who usually vote Nationals are likely to vote KAP instead… and are likely to follow KAP HTV.
As a number of people have pointed out, most phone polling is done via landline phones, use of which has been in decline in the last three years, resulting in a significant chunk of the electorate not being polled. I make no claim about the bias resulting from it, as I don’t know what impact it has; but it does reduce confidence in the results (and it’s systematic, meaning you’d see it on all of the phone polls, unless they call mobiles).
What’s more, as a result of things like the senate “dirty deals”, a lot of people are becoming more aware of options of figuring out their own preferences, even in the senate. There are websites out there set up to make it easier for people.
Every one of these effects would be expected to be independent of the specific polling company, and thus would result in an inherent bias in all polling. And they invalidate the assumption of past preference flows being representative of 2013 preference flows.
In short, we just don’t know.
By the way, people said Essential leans to Labor. Essential were the only ones to be “accurate” in 2010 in their final poll before the election.
Glen, the factors you point to in trying to explain why this time a one-point difference (as it has averaged throughout the campaign) from one pollster better explains reality just doesn’t wash. There is nothing unique about those circumstances. And the KAP/PUP phenomenon is highly unlikely to mean Labor will get enough more of of the non-three-party vote to actually amount to a full percentage point (or to even get more at all). That’s compounded by the fact that only PUP, of those two, has candidate everywhere, and that even with those candidates, both parties are Queensland-based. Labor took 78.84% of the Greens vote and 41.74% of the minor party vote in 2010.
Let’s just take the last real multi-mode. Assuming that Labor gets roughly the same share of Greens preferences as last time (about 79%), Labor would need an increase in its share of Other preferences from about 41% to 52% to account for a one-point change in TPP. If you look at the more recent phone+internet poll and believe that polls understate minor parties (I think the opposite is usually true) and so its 13.5% Other result (3.5pp higher than the next highest and 5.5 higher than the BT estimate) is plausible, then Labor only needs to take just under 50% of the Other vote. I find it hard to believe that changes in HTV cards and KAP will provide that big a boost or that PUP will do anything but work against Labor’s share of the Other vote.
Glen, didn’t Newspoll say 50.2 for the last election? That was 1/10th of a percentage point off the actual result.
Morgieb – my apologies, their very last one was indeed 50.2. I was remembering the ones before it.
I just looked into it a bit further; Newspoll was closest on 2PP, but they were well off on primary votes. Essential were the most accurate on primary vote.
http://www.smh.com.au/federal-politics/political-opinion/winner-declared-in-pollsters-election-prediction-race-20100924-15poh.html
It’s definitely an interesting thing to look into.
Labor is getting smashed in Newspoll out tonight. Numbers to be revealed after 10pm I believe.
I have a non-poll-related question, perhaps a silly one: If the campaign period is typically 33 days, why do the parties “launch” their campaigns halfway or more through? A similar question could be asked about candidates, who seem to wait until almost halfway. Obviously what they want is an opportunity to grab media attention, but why weren’t they mocked mercilessly when this started to nip it in the bud?
DB are you getting that from the SkyNews reports or have you actually seen the newspoll numbers internally?
Poll Bludger has tomorrow’s Newspoll as 54-46 to the Coalition.
Comments are closed.