Losing The Republic Like It's 1999: The Polling

The current Voice referendum with its rapidly declining support has brought back memories of referendums past.  Surprisingly, given that it's our most recent referendum, there has been little information available about the polling trajectory of the 1999 Republic defeat.  The received wisdom is that support for the Republic started off very high and ended up low, as happened with several other referendums (especially the 1988 quadruple failure) and as is so far happening with the Voice.

It turns out to be not so simple.  The story of the 1999 Republic referendum polling is one where only the last few months of data are all that usable, and in public polls at least, Yes was never all that much in front.

A summary of Republic polling is available in one paper available online but I thought, surely there was more?  I could not find any usable online archive of Republic referendum polling but what I did find, to my surprise, was some polling commentary in an article I published in Togatus April 1999.  1999 model me - not then a polling analyst so please cut some slack for misusing "push-poll" for a seriously bad polling question - wrote this:

"Poll results are confusing. In the last week of January the Age/AC Nielsen showed a national 41% yes vote, but Newspoll showed 58%.[*] The latter was a virtual push-poll because it included the statement 'this will most likely mean that the head of state will not be a politician' and ARM won't get even that simple message through to a thick republican public that mistakes an extra election with people-power."

[* 59% actually. This Newspoll wording was said to have been sourced from the ARM, and was roundly and rightly condemned by opponents as out of step with other polling at the time.]

In fact, question wording was a major issue in 1999 polling, and a case where public polling itself influenced the ballot paper design for the better.

Wonk Central: The Track Record Of Last-Election Preferences

NOTE: Updates for elections from 2016 on appear at the bottom of this article

--------------------------------------------------------------------------------------------------------------

Advance Summary

1. This article uses historic data to examine how good preference flows from the previous election are at predicting preference flows at any given federal election.

2. Last-election preferences have been an industry standard following various failures by and problems with the respondent-preference method, especially its total failure at the 2004 election.

3. Using a model of minor-party breakdowns similar to that used by most pollsters, this article looks at expected versus actual preference flows at all federal elections since 1955.

4. The quality of data available on preference flows is better for elections from 1983 onwards.

5.  Since 1983, last-election preferences have predicted preference flows remarkably well at most elections.

6. However, last-election preferences substantially understated the flow to Labor at two elections in this time: 1990 and 2013.

7. The historical record of last-election federal preferences is so strong that claims that last-election preferences in polling are wrong should generally be treated with great caution (including now).

Could A UK-Style Mass Pollster Fail Happen In Australia?

Advance Summary

1. Recently all major final polls and the vast majority of analysts failed to predict the UK election correctly.

2. Random sampling error, "shy Tory effect" and late swing probably did not make major contributions to this outcome.

3. More likely causes include herding (although this has not been conclusively shown to have happened), the difficulty of estimating turnout, and the abundance of essentially non-random "panel polling" methods.

4. Australian elections are easier to poll for because almost everyone votes and because there is very little tactical voting in the Lower House.

5. Despite this there is a greater risk of high average polling errors at the next federal election because of a rapid turnover of polling ownership and methods, which will make it more difficult for pollsters to detect issues with their methods.

6. Potential for Labor support to be soft even in very late polling also appears at this very early stage to be likely to be higher than normal.

Poll Roundup: 2014 Year Review

2PP aggregate: 53.7 to ALP (+0.3 since last week)
ALP would easily win election "held now"

We're just about at the end of the federal polling season, an end which for the government can't possibly come soon enough.  If last year's polling is any guide, we may get one more ReachTEL at the end of this week and then that would likely be it until Morgan and Essential started up again in mid-January with the rest following a few weeks later. I thought I'd summarise some stats about this polling year, but first a quick look at this week's polls so far.

I should note first that a significant event this week - the Martin Place siege - could affect voting intention to some degree, and that incident was not reflected in any of this week's polling.  The past history of this government is that its polling improves temporarily and modestly when national security is in the spotlight.  Because of the scarcity of polls over the holiday season there may not be a lot to look at by way of evidence, and it remains to be seen whether this disturbing incident will be perceived as a terror issue as much as a run-of-the-mill failure of the domestic justice system. 

Poll Roundup: A Brief History Of Disaster Bounces

2PP Aggregate: 53.1 to ALP (-0.6 since last week)

"I actually think Tony Abbott is doing us proud.  I didn't think he would, I didn't think he had it in him, but he's put his foot down and now other countries are supporting him".

The above was one of a number of vox-pop style voices in a Tasmanian ABC radio segment about reactions to the MH-17 air disaster.  Interviewees, all or nearly all of them female, spoke about the way the disaster is being seen in the community and the issues involved in talking about it to children.

MH-17 coverage is inescapable in media of just about any kind, and some of the more lurid excesses of disaster-porn (Daily Telegraph and Herald Sun, that would be you) have just about had me reaching for the JG Ballard collection in disgust.  But what effect does this sort of thing have on polling and perceptions of the government?

When I put to various left-leaning online political circles the idea that this might see a rather large improvement in the Abbott government's fortunes, the consensus reaction was pretty sceptical: no bounce, or if any bounce lost in the noise.  After three polls out this week, we're seeing what looks like some movement back to the Coalition, but it's modest so far and far from conclusively real.  Untangling what is going on is made more difficult by the government having succeeded in repealing the carbon tax in the same polling cycle.

Victorian Liberals: Going, Going ...

This is an overview of the state of play in the leadup to the Victorian election, prompted by recent polling.  I expect to make more detailed modelling attempts much closer to that poll.  In summary, this article argues that the Baillieu-turned-Napthine government's first-term status provides no argument against the reasons why it could lose.  While the defeat of this government is not yet a done deal, psephologically, it has at least one foot in the grave.  There is still time to climb out, but it won't be easy.

One of the themes I've been following on this site is the way in which federal politics contaminates state voting intention.  As noted in Is Campbell Newman Actually In Trouble? life gets much harder for state governments when they have a party of their own persuasion in power in Canberra, especially if the federal party isn't polling well.  This is also well covered in Peter Brent's current piece, Dog Days For State Conservatives.  Of the four conservative state governments elected while Labor was in office federally, all are showing signs of wear and tear.  We still have astonishingly sparse state polling from Western Australia since Colin Barnett polled a very poor personal result, but it wouldn't be at all surprising to see Labor in the lead there soon.  The NSW polling picture is still re-sorting itself after the shock loss of a Premier, and in Queensland the LNP have only a narrow lead (Queensland polling updates added) and are still struggling greatly to build their vote share to the level needed to make their unpopular Premier's own seat safe.

Budget Blowout ... In Coalition Polling

Scroll to bottom for latest updates - updated through May and first week of June

Current aggregate: 54.2 to Labor (+0.3 since last week, +2.9 in five weeks)
Tracking graph (smoothed)

Advance Summary

After holding a slim and by no means election-winning polling lead for five months from November 2013 to April 2014, the Labor Opposition moved into a clear and widening polling lead in the weeks surrounding the 2014 Budget, and would easily win a hypothetical election "held now".  The polling position of the Coalition is roughly comparable to the two worst slumps of John Howard's first term, but many federal governments have recovered from similar, and in at least four cases worse, polling situations.

The federal budget is among the worst received in the last 27 years, second only to the 1993 "horror budget" in terms of its perceived impact on poll respondents and providing the worst ratings since the early 1990s in terms of perceptions of its impact on the economy.  Tony Abbott's personal ratings are among the worst recorded in one pollster's history (Nielsen), while in another's (Newspoll) they do not stand out so much, but are nonetheless the worst results he has recorded since late 2012.

Nielsen: Greens 17: What Is This?

This week's AC Nielsen poll has created a bit of a flap with a primary reading for the Greens of 17%, and that is not the only unusual thing about it.  This result is a national record - the highest vote polled by the Greens in any reputable national opinion poll.  Ever!  Their previous highest results were recorded during a purple patch surrounding the 2010 election - 16% in Newspoll in late May that year, 15% in Nielsen the week later, 15% in Galaxy in late July, and two 15s in Morgan that September. 

The state breakdowns in the Nielsen (in a preliminary version the PUP row was mistakenly labelled "Independent") are especially interesting because they show the Greens on 27% in WA, compared to only 20% for the ALP, despite the two-party-preferred in that state being 52:48 to Coalition.  I've modelled the implied 2PP from the Senate poll as 53:47, so 52:48 in WA is about what we'd expect. Of course, the sample size for WA is tiny, so we shouldn't take too much notice of the figure.  All the same, might such things as a national Greens vote of 17% really be true?

Poll Roundup: National Polling All Over The Place

Poll Aggregate: 51.2% to Labor (update: 50.4% on 18 March)
Labor maintains narrow lead held since December

Nielsen 52-48 to Coalition? Newspoll 54-46 to Labor?  These two results just one week apart - what could it mean?

It's about time to have another of my irregularly scheduled roundups of the national polling picture and how the not quite so new Abbott government is travelling.  My last federal roundup of sorts was a month back in the Griffith article.  At the time Labor seemed to be starting to drop off a little, and the lukewarm by-election result seemed consistent with that.

Shortly afterwards we had the 52-48 to Coalition from Nielsen.  At the time some dismissed it as a rogue, but it also seemed it could be a harbinger of a return to a Coalition lead.  There was some support for this from a Morgan which gave Labor a wafer-thin 50.5:49.5 lead by last-election preferences (equivalent to a thin Coalition lead given the apparent house effect of Morgan's methods).  Also, two weeks of 51:49 to Labor from Essential proved little, since we know Essential can be contrary to, or lag, the trend of other polling.

Federal 2PP Aggregate Methods (44th Parliament)

This post describes the revised aggregation methods used to produce the two-party preferred aggregate (unofficially nicknamed "Aggedor") that appears in the sidebar of this site.  Although slightly more detailed than the version used before the 2013 election (which predicted the election 2PP of 53.5 as accurately as possible, albeit with quite some luck) or the transitional model used in the months after the election, it is nonetheless designed to be mathematically simple and computable quickly by hand at any time.  I've also tried to increase the proportion of the model that is objectively defined, though there will still be subjective judgement calls from time to time.

The aggregate is essentially a weighted average of two-party preferred polling derived from all recent polls that are considered of sufficient standard.  2PP figures from each poll (see "The 2PP score that is used" below) are multiplied by various weightings based on their recency, accuracy and other issues, and then the sum of the multiplied poll scores is divided by the sum of the weightings.

Pollsters Included:

Polls by the following pollsters are included when available: Galaxy, Newspoll, Nielsen, ReachTEL, Morgan, Essential, AMR, Ipsos.  Other pollsters may be added subject to their completion of three polls of national 2PP voting intention by the same methods, and adequate methods documentation.  Even if a poll is by an included pollster, the poll is not included if it is an "internal poll" (defined as a poll commissioned by a political party, union, business or lobby group).

2013 Federal Election: Best And Worst Pollsters

Ideally I would have written this article a few weeks after the election, while debates about the quality of polling before the election were still fresh in everyone's minds.  It seems a world away now but earlier this year there were actually plenty of people arguing that the day of the landline was done, that online and mobile polling were the way to go and that Palmer and Katter supporters would preference Labor in droves thus making the election much, much closer than all those pesky polls said.

But now ...

This was, in general, a pretty good election for the pollsters (most of them anyway) and a bad election for those who wanted to outguess them.  

For basic information about the different polls and their methods, and a lot of general stuff about polling methods and debate surrounding them, you may find the recently updated Field Guide To Australian Opinion Polls useful. 

I'll consider pollster quality in three categories:

Early Abbott Era Polling Roundup

(Note: an update has been added at the bottom following the surprising first Nielsen result and other signs of narrowing in late November.  Another has been added for the release of official 2PP preferences.)

We are now just over two months into the lifespan of the first Abbott government and it is time for a brief summary of its polling history so far, and that of its leader. 

The following polls have been published since the 2013 election: nine Essentials (one of them including a week of pre-election data), four Morgans, two Newspolls and one and a half ReachTELs.  (The half ReachTEL was a poll of NSW and Victorian voters only, that canvassed voting intention along with Shorten-vs-Albanese ALP leader polling.)

Two-party preferred

Compiling a meaningful two-party preferred picture from this polling has not been easy, for the following reasons:

* The performance of Essential and Morgan in the period between the return of Kevin Rudd and the 2013 election.  Essential tracked in a contrary manner to other pollsters (below them for Labor after Rudd returned but then above them for Labor as the election drew closer).  Morgan (using preferences modelled by the last election) initially tracked favourably for Labor but once the election was called tracked neutrally.

* A method change by Morgan to drop online surveying from their most recent survey. 

2013 Federal Election Late Counting - House Of Reps

This post will follow late counting in the House of Representatives. The Current State of Play will always appear at the top of the screen, followed by Fairfax updates, some general comments from the early history of this article, and the newest non-Fairfax updates at the top of the section below that.

Current State Of Play (This will be changed when status changes)

Last updated 31 Oct (Fairfax)

Election result is 90 Coalition 55 Labor 5 Other (2 Ind 1 Green 1 KAP 1 PUP)

Fairfax: What Happens Now, and Past Close Finishes

31 Oct: It's over! (At least to the point the bookies pay out at; it remains to be seen if the LNP want to risk antagonising the PUP leader by taking this to Disputed Returns.) Clive Palmer will be declared the winner of Fairfax by 53 votes at the end of a massive struggle over an electorate that was not deemed worthy of a single published seat poll.

That officially concludes this thread, which has been the most visited thread on the site so far.  Thankyou all for your interest.

Saturday (Final!) Federal Election Projection Model and Seat Betting Roundup

Note: I will be at themercury.com.au on election night.  A pointer to the coverage will be here.

PROJECTION 10 AM SATURDAY(FINAL)
COALITION WINS ELECTION (probability >99%)
SEAT TOTAL PROJECTION 95 COALITION 52 LABOR 3 REST
(This is an approximate projection only.)

Finally the time is reached for pollsters to release their final polls, modellers to release final projections and politicians to make final attempts to woo the "undecided vote" (which is usually, in real terms, a lot smaller than a lot of people think.)

Late And Final Polls

Essential was first of the regular pollsters to release what seems to be its final result, a 52:48 based on 1035 interviews from Sep 1-4.  If the result for Labor is benign or even an improvement compared to current polling by other sources, Essential will look very good, but if election day is a rout then there will be questions asked about going too early, using a smallish sample and the pollster's trending behaviour relative to others. Galaxy has released what seems to be a pretty safe 53:47, again showing statistical properties more akin to an aggregate than a poll; most likely if they're wrong at all it won't be by much (I am unsure if that's their final offer.)  ReachTEL is providing useful daily tracking and finals from Newspoll, Nielsen and others will arrive later today.  Morgan has released a multi-mode with respondent preferences at 53.5:46.5 (54:46 by last election); they will have a final poll tonight. (Update: this turned out to be just extra data for 54.5: 46.5)

Poll Roundup and Seat Betting Watch: New Dawn Fades Edition (August 27)

2PP Aggregate (Tuesday 27 August): 52.3 TO COALITION (+0.3 since last week)
Individual Seat Betting: Labor favourites in 59.5 seats (-1.5, Eden-Monaro, Brand becomes tossup)
Seat Total Market: Labor 61 seats (-2) (This figure is probably slightly skewed by longshot bias.)

This is week nine in a regular weekly series in the leadup to the federal election.  Week eight was here and through it you can click back to the previous weeks.  Or just click the "betting" label at the bottom.  As stated before, the aim of this exercise is not to claim that seat betting markets have predictive value, but to test whether they do, and to see which of the markets and the aggregated polls see the ultimate outcome of the election first.

This Week's Polls

It was a relatively quiet weekend for polling compared to the previous week's six-poll deluge.  Nielsen and Newspoll both returned 53-47 to the Coalition.  Roy Morgan returned 52.5-47.5 by last-election preferences.  Essential returned a second consecutive 50-50, but this one was off the back of a two-point fall in the ALP primary and a three-point rise in the Greens primary. My aggregate has gone to 52.3.  At the moment, it has one of the more benign readings for Labor doing the rounds - aggregates that benchmark solely off the last election are about a point worse, and those that handle sample size issues in a more sophisticated way, or weight heavily for accuracy of particular polls, are also showing slightly worse readings for Labor.  An aggregate of all the aggregators listed in Poll Bludger's excellent review of the state of things here comes out to 52.85.  My Newspoll rolling average, which I use heavily in historical modelling, is at 53.

Poll Roundup and Seat Betting Watch: Scruff Of The Neck Edition (August 13)

2PP Aggregate (Tuesday 13 August): 51.5 TO COALITION (+0.7 since last week)
Individual Seat Betting: Labor favourites in 63 seats (-2, Forde to Labor; Banks, Reid, Brisbane to Coalition)
Seat Total Market: Labor 65 seats (-2) (This figure may be slightly skewed by longshot bias; see below.)

This is week seven in a regular weekly series in the leadup to the federal election.  Week six was here and through it you can click back to the previous weeks.  Or just click the "betting" label at the bottom.  As stated before, the aim of this exercise is not to claim that seat betting markets have predictive value, but to test whether they do, and to see which of the markets and the aggregated polls see the ultimate outcome of the election first.

Things aren't looking too good for the Government after a few weeks of apparent competitiveness.  The polls say they're losing at the moment, the punters strongly think they'll lose, the voters think they'll lose and the modellers think they'll lose.   Still left expecting a Labor victory are really just the true believers, the Abbott-haters (and not even all of those), the Sawford formula devotees and Peter Brent.  There's a general perception that the Coalition now has this election by the scruff of the neck, that the first month of the Rudd comeback polling was partly bounce and that the government won't be going up from here and may indeed be going down further. Is it really quite that hopeless for Labor?

A Field Guide To Australian Opinion Pollsters

As of September 2016 this is an archived old edition that is no longer being updated.  For the 45th Parliament (2016-) edition go here.


Introduction

There are a lot of polls about in Australia these days.  But how do they all work, which ones have runs on the board and which ones can you trust the most? 

With an increased number of pollsters now polling in Australia, it's been suggested that an article describing what is known of each pollster and its strengths and weaknesses might be useful to many people.  So I thought I'd put one up, and edit it over time as the need arises.  (Now edited post-2013 election.  Last major edit 30 Nov 2013. Last minor edits 21 June 2016.)

The gold standard for success for an opinion pollster is that its polls at election time get the result as close to right as possible.  However, many of the new pollsters are little-tested against actual elections, and getting a specific election right is a combination of skill and luck.  In elections where there is a swing on the last day or two of the campaign, a pollster that is actually not polling correctly may have its errors cancelled out by the swing, and hence record a lucky hit.  There is more to being a good pollster than just getting it right at election time - a good pollster should also provide useful data between elections and do so using well-designed questions that are easy to interpret.  And a pollster should also present their data in a way that makes sense and isn't misleading or confusing.

Bob Ellis: Embarrassment To The Left

Bob Ellis is the Unskewed Polls of Australian politics.

A digression: once, I used to feel some sympathy for Ellis.  Some may recall Bob Ellis cost his publishers $277,000 with basically a single multiply false sentence in the book "Goodbye Jerusalem", which defamed Tony Abbott, his wife, and Tanya and Peter Costello.   When the full facts about Tony Abbott's abandoned "son" came to light, I thought that the discrepancy between Abbott's former personal life and Abbott's religious-morality attitudes as a Howard Government minister said far more to Abbott's detriment than any scuttlebutt falsely claiming he was sexually seduced into the Liberal Party ever could.  There are some small things in that case that might have gone down differently had Abbott's full biography been known when it was heard. 

However, the sympathy I had for Ellis from that has been wiped away many times over by his recent series of comically bad attacks on mainstream opinion polling in Australia.  Quite aside from his the defamatory and tinfoil-hat nature of much of his work, a big problem with Ellis's comments is that what he writes about opinion polls is riddled with easily avoidable factual errors.

Poll Roundup and Seat Betting Watch July 16

2PP Aggregate (Tuesday 16 July): 50.0 TIED (+0.2 for ALP since last week)
Individual Seat Betting: Labor favourites in 63 seats (+2 this week - Lingiari, Brisbane)
Seat Total Market: Labor 68 seats (+2)

This is instalment three in what seems to be becoming a regular weekly series in the leadup to the election.  The first Seat Betting Watch is here and last week's Poll Roundup and Seat Betting Watch is here.  As stated before, the aim of this exercise is not to claim that seat betting markets have predictive value, but to test whether they do, and to see which of the markets and the aggregated polls see the ultimate outcome of the election first.


This Week's Polls

As well as the 51-49 to Labor from AMR late last week, we've now had a 50-50 from Nielsen, a 52-48 to the Coalition from Essential and a 51.5-48.5 to Labor by last-election preferences from Morgan Multi-Mode.  (I did cop an "AGGGGHHH!!!!" and a "stupid" on Twitter for continuing to prefer last-election preferences to respondent-allocated but I'll believe KAP voters have turned into lefties only when data show they have.)

The net impact of these on my aggregate has been small and I currently have the two parties dead level on two-party-preferred vote share.  Essential and Morgan don't seem to be behaving the same way compared to other pollsters as they did before Rudd was reinstalled, and Mark the Ballot has a nice wrap of what a pain in the neck this all is for modellers here. (Also see the comment by Julian King on that article.)  Fortunately, the probably false assumption that neither has a significant house effect results in them largely cancelling each other out in my very simple model, but it's a bigger problem for those who take Morgan's large sample size into account.  It may turn out that either of these pollsters has struck it lucky and all the rest are too high or too low, but I'll keep assuming that that isn't the case.  Something to keep an eye on here is that Rudd is doing very well with young voters, and young voters are the most difficult to sample accurately for conventional polls, even with scaling.

Federal 2PP estimate feature added

 (Note: This article documents the experimental phase of the aggregate, which ran from February to September 2013.  For transitional arrangements post the election see here.)

I've just added a subjective two-party-preferred vote (2PP) estimate feature on the sidebar.  The reason I have added this feature is that I commonly get involved in discussions on various sites in which someone is making claims about the likely state of the national 2PP that aren't even remotely credible, or just getting confused about all the different polls and the strange range of values they spit out.  I think a fair few people will from time to time be interested in my view of where things are at, especially when having those kinds of debates while I'm not around.  There are some handy formal aggregators about, but they often take a few days to update and I often find my view a little out from theirs (and usually somewhere in the middle of them all), typically because everyone has slightly different views on the best underlying assumptions.  When I can, I'll be aiming to update this estimate quickly.

The 2PP estimate is not a fully formalised aggregator and is not a scientific test. I'm not at the point of being ready to attempt something like that yet, and while I have a long-running Newspoll rolling-average of sorts for historical comparison purposes, I've only recently developed an interest in trying to gauge the picture across all federal pollsters.  It's just my hopefully informed opinion on the basis of an informal and at this stage loosely defined aggregation process.  The rough assumptions I make in thinking about the national 2PP are as follows.  (Note: Article has been edited to give the current version.  Legacy text appears at the bottom so people can see how this model has developed.)