Polls Archive

10 days to go: Senate poll strong for Greens

Roy Morgan has released their first polling of the Senate race, showing a very strong vote for the Greens.

The poll shows the ALP on 40%, the Coalition on 36%, and the Greens on 15.5%. On a state level, the Greens are polling strongly in all states, the best results being 18% in Western Australia and 17% in New South Wales, followed by 16.5% in South Australia, 14% in Victoria and 13% in Queensland, all enough to elect a Senator with a small amount of preferences in Victoria and Queensland.

In Tasmania, the Greens are on 21.5%, easily enough to elect one Senator but not enough to give them a chance of a second. In the ACT, Morgan’s poll has the Greens even with the Liberals on 27%, which would give them the seat on Labor preferences.

Overall, the result would produce a Senate with ten Greens, 33-34 Labor senators, and 31-32 Coalition senators, with Family First’s Steve Fielding losing his seat.

It has been widely argued that it is impossible to do Senate-specific polling. Many make decisions on how they vote based on the House of Representatives, and designs of ballot papers make the decision-making process very different. Having said that, there is still value in examining the effectiveness of this poll.

Despite criticisms in the past, Morgan’s Senate polls in 2007 saw the Greens bounce around between 8.5% and 9.5%, before settling on 9% in October 2007: almost exactly what the Greens polled in the Senate in 2007.

When examining the state breakdowns, they follow a different pattern to recent state breakdowns produced by Nielsen, although the major party votes follow similar patterns.

In terms of the major party vote, the ALP is up in Tasmania, about even in Victoria and South Australia, and down in New South Wales, Queensland and Victoria.

In terms of the Greens vote, recent Nielsen polls have had the Greens polling highest in Victoria and Western Australia. I haven’t noticed a massive difference in the Victorian and New South Wales campaigns from the Greens, although the ┬áVictorian campaign has focused more on the race in Melbourne than the Greens NSW have on Sydney. Even still, the consensus has been that, for a number of reasons, the Victorians are expected to achieve a higher vote. We’ll have to wait and see if the Morgan poll is right, and Lee Rhiannon is going to easily win election.

This poll also predicts the Greens winning the seat currently held by the Liberals in the ACT. The Greens vote level is similar to a poll commissioned by the ACT Greens that put Lin Hatfield-Dodds on 26%.

While overall the figures are extremely positive for the Greens and should be taken with a grain of salt, it’s worth noting that the 15% for the Greens is not that far above the 13% received in this week’s House of Representatives polls. If you assume the Greens will poll slightly higher in the Senate, 15% is not that far off.

Galaxy/Nielsen: 52-48 to Labor, Greens on 14/13

I don’t usually post about the regular polls that we will see more of over the next few weeks as we plunge into the election, but I’ll make an exception for the two polls coming out overnight.

We have had two polls in the last few hours. The Galaxy poll in the Courier-Mail and the Neilsen poll in Fairfax newspapers. Both have Julia Gillard on 52% on two-party preferred teams, and both have the Greens recovering votes, up from 11 to 14 in Galaxy and up from 8 to 13 in Neilsen.

It is now worth noting that we now have a series of polls since Gillard took over as Prime Minister. All have the ALP in a winning position for the election, but with the exception of the weekend’s Morgan poll which had the ALP on 56.5%, none of them have the ALP on any more than 55%.

While they are a slight improvement over Kevin Rudd’s performance, overall the ALP’s chances of re-election have only really improved slightly. Kevin Rudd’s polls put him in a position where he was the favourite to win. After the last week’s events around asylum seeker policies, you would have to say that Julia Gillard, while the favourite to win the election, could very easily lose to Tony Abbott if things don’t go her way.

We now also have the first poll showing a recovery in support for the Greens after the Green vote fell when Gillard became Prime Minister. It isn’t at all surprising that the Greens would be recovering considering recent developments in climate change, asylum seeker and mining tax policy, but does indicate that the appeal of an atheist, childless woman as Prime Minister can only obscure the rightward drift of the Labor government for so long.

Having said all of that, it does appear that not a great deal has changed in terms of polling since Kevin Rudd was deposed. The ALP has a small but election-winning lead over the Liberal Party, with the Greens on track for an increased vote, but much of that appears soft and willing to consider switching back to Labor. You’d have to say that, based solely on polling, the overthrow of Kevin Rudd was one of the biggest over-reactions in recent political history.

More psephojunk from the Australian

The entirely forgotten former Labor senator John Black has published a very strange and stupid piece of electoral analysis in yesterday’s Australian, basically arguing that Greens voters are richer than Liberals, therefore they won’t support the mining tax because it will hurt their share portfolios. This will apparently cost the ALP the election because every single new Greens voter came from the ALP, but they will be preferencing the Liberals en masse because of the ALP’s tax on mining.

It’s difficult to decide which part of Black’s stupidity is the worst. When you get past the snide insulting of Greens voters (I don’t know what evidence he has that Greens voters don’t study maths or economics – my experience certainly doesn’t indicate that).

Black seems to base his entire thesis on the idea that Greens voters are the wealthiest group of Australians. It’s true that the core Greens constituency consists of middle-class Australians who generally are doing relatively well, although it also includes plenty of young people and students who have low incomes.

Before trying to compare political parties, you have to recognise that major parties are made up of a number of constituencies each larger than the Greens. Labor voters include working class voters with lower educations and inner-city progressives with degrees (who are very similar demographically to a lot of Greens voters). The Coalition draws support from wealthy, highly-educated voters as well as working class voters in the fringes of the big cities and Nationals voters in rural areas without much of an education. If you average those out, the Greens may come out with the highest average (although I doubt it). It’s far too simple to say “rich people vote Green”, and is obviously absurd.

There is a variety of reasons why Greens voters are seen as relatively wealthy. Greens voters tend to be young, so they don’t tend to include as many retired voters with depressed incomes (in contrast to the Coalition). Female Greens voters would be much more likely to be employed, which also would increase incomes. Indeed, I have seen evidence that, amongst highly-educated people, Greens voters tend to be less wealthy. Greens voters are wealthier mainly because they have higher education levels.

Black argues that the Greens are serving in a similar role as the Democratic Labor Party played in the 1950s and 1960s, when it directed the preferences of Catholics away from Labor and to the Liberals. It’s a really strange comparison. First of all, the DLP actively worked to prevent the ALP from forming government, directing preferences to Liberal candidates and helping the Liberals stay in power continually for 23 years. In contrast, the Greens very rarely preference the Coalition and their preferences were central to the election of the Rudd government in 2007.

The political base of the two parties is also extremely different. The DLP’s positioning on the right wing of the ALP as opponents too communists and the ALP’s left meant that they sat in the centre of the political spectrum. The Greens instead sit to the left of the ALP. It seems very strange to think that the Greens could act as a bridge to ease the transition of voters from Labor to Liberal.

Black’s thesis appears to be based on the assumption that Greens voters are motivated by their self-interest, and they see the health of mining profits to be in their self-interest. This seems like a very courageous assumption. Greens voters are clearly strongly motivated by environmental issues and social justice. While Greens voters are usually middle-class, the party has built a base through campaigning for public health and education in preference to private alternatives and campaigning for strong action on climate change.

Indeed, the Greens have been fiercely critical of the mining industry and advocated for the gradual shutdown of the coal industry. Black figures that rich Greens voters are worried about their share portfolio and superannuation if the federal government imposes a tax on the mining industry, despite Greens voters strongly supporting action on climate change which would hit the resource extraction industry hardest.

There is no evidence at all that Greens voters oppose the RSPT. Bob Brown today has criticised the government’s policy for subsidising mining, and that position seems to be in line with where Greens voters would stand. So is there any evidence at all that Greens voters are being pushed away from preferencing Labor because of Rudd’s policies on mining taxes? I can’t find any. It is probably reasonable to assume that most new Greens voters are coming from the ALP, but it is also reasonable to assume that most of those voters will return to Labor in preferences, despite their gross disillusionment with the Rudd government.

If Labor is to lose power in 2010, it will be due to voters in the centre switching to the Liberals, it won’t be due to left-wing voters deciding to preference Tony Abbott because Kevin Rudd is too hard on the mining industry!

Tags: ,

Beware the leaked Labor polling

Ever since the ALP lost the 2002 Cunningham by-election to the Greens with a large protest vote, they have finessed the art of expectations management through the leaking of supposed internal party polling data. The latest example came yesterday when the Sun-Herald ran a story reporting that leaked Labor polling puts the ALP on only 32% of the two-party preferred vote in Penrith in the lead up to the June 19 by-election.

This story indicates a couple of things: firstly, it indicates how desperate the ALP is. It seems very implausible that the ALP is polling so badly, but it indicates that the ALP is looking to use this story to scare Labor voters back. Secondly, it shows how eager some journalists are to be spoonfed a story by spin doctors if it creates some political drama. It sells newspapers to publish a front-page story reporting such catastrophic polling figures, while burying the key facts that such a poll has zero scientific credibility due to its source.

Considering the ALP’s actions in leaking the poll, it suggests we are heading towards a result comparable to the 2008 Ryde by-election, with Labor losing on a large swing. Which we already could have guessed.

UK 2010 – Election Day prediction

UK voters are now casting their votes in the 2010 general election. Over the last few days we have seen a clear trend in polls in the UK: the Conservatives are clearly in first place around 35%, with the Lib Dems and Labour in a statistical dead heat in second place around 28%.

While these figures are quite clear and consistent amongst all the polls, it is extremely difficult to make predictions. Polls in the past have overestimated Labour’s support, which would suggest we are on track for the Lib Dems to come second. On the other hand, that tradition took place in an era when voting Labour was popular and cool, while voting Conservative was not. It is possible that such a trend could be reversed, and “shy Tories” could be replaced by “shy Labour”.

The biggest question, however, is how the vote numbers translate into seats. Most of the British media has been using the “Uniform National Swing” model to produce figures for how they expect the numbers to go, and expect a variation on this model to be used by the BBC to extrapolate their exit poll into a result. Having said that, you would have to think that with such massive turmoil in the polls and shifts in support and the rise of the Lib Dems, you would expect the swing to be very non-uniform, and in ways that could effect it. I would expect that the result would deviate away from UNS in such a way that sees Labour lose more seats to both the Tories and the Lib Dems.

This has been reflected in the model used by US polling blog FiveThirtyEight, who have developed a model where they make interpretations about which voters are switching between parties and use a variation on proportional swing. While the UNS model used by FiveThirtyEight sees a 36-28 gap between Tory and Labour leave the Tories with only ten seats more (and so far short of a majority that only the Lib Dems could deliver a majority), the FiveThirtyEight model on similar figures sees Labour lose sixty more seats: fourty to the Conservatives and twenty to the Lib Dems.

Despite my earlier refusal, I think it’s worth me taking a punt at a prediction. I won’t make detailed seat-by-seat, or even region-by-region picks, but I will make a few broad statements.

First of all, I think the polls are mostly right, but if anything the Lib Dems could be slightly underestimated, which could see them overtake Labour, but I give Labour a 60% chance of coming second. Apart from that, the voting figures of 36/28/28 will be roughly reflected. UK polls have improved tremendously and the final polls are all very close to each other this time around.

So here’s my prediction. Let’s start with the small parties. I predict that the Greens will win Brighton Pavilion, where they will gain the boost received elsewhere by the Lib Dems, who haven’t shown much signs of campaigning in the seat, so shouldn’t siphon too many votes away from Labour and Green.

I also predict that George Galloway will pull off a win in Poplar & Limehouse, but that Labour will regain his seat of Bethnal Green & Bow, but I reckon there could be a chance that the Respect candidate could retain that seat, giving the party two seats in eastern London.

Plaid Cymru will retain their three seats, despite the redistribution reducing their numbers to two. I predict the SNP will gain seven seats. Without trying to name them, their surge of 2007-8 has receded and Labour appears to be mostly maintaining its position in Scotland, with the small protest vote against them going to the Lib Dems. Having said that, the sole Conservative MP could well lose his seat, leaving them with no seats there again.

I also predict that Sinn Fein and the SDLP will retain their seats (5 and 3 respectively), but I won’t try and make predictions about how the ten unionists seats will be split between the DUP and the UUP/Conservative alliance.

I also predict that Speaker John Bercow and the two independents will retain their seats.

As for the majors, I won’t pick exact numbers, but the Conservatives will get just short of 320 (my numbers came out as 319), Labour just short of 200 and the Lib Dems around 100, possibly slightly over.

The consequences of this result would be a hung parliament, but not one where the Lib Dems are the key powerbrokers. With the Conservatives so close to a majority the option of a Labour/Lib Dem pact is not available. In addition, the Tories would have a number of options, including the Lib Dems, the Celtic nationalists and the Northern Ireland unionists. They could stitch up a deal with the DUP, the Lib Dems, or attempt to go it alone, and wouldn’t have too much trouble getting support on each issue. You would likely see Cameron govern moderately and modestly for a year or two before calling another election when he sees the opportunity to gain a majority.

This is the scenario where the order of votes between Labour and Lib Dem will become important. While I predict the Lib Dems to win around 100 seats, you would expect a lot more seats to become marginals on the Lib Dem target list, seats where the party was not expected to win but had performed strongly. If Labour manages to come second I expect they will pull together and unite to prevent Cameron’s government from functioning with a working majority. But if the Lib Dems manage to come second they will be put in a position where an election a year or two down the track could see them challenge Labour in many more seats and eclipse them as the major force on the left in British politics.

Polls close at 10pm in the UK (7am Australian Eastern Standard Time) and I will be up to blog from then on. At that time we should see the joint BBC/ITN/Sky exit poll, and begin to see results within two hours.

UK 2010 – the final week

Voters in the United Kingdom go to the polls this Thursday, and the race remains one of the most fascinating in recent history.

Lib Dem leader Nick Clegg has managed to largely hold on to the support he gained following the first debate, although commentators disagreed about whether Clegg or Conservative leader David Cameron won the second and third debates.

The last few days of the campaign haven’t gone well for Gordon Brown, with it reaching its low point on Friday, when Brown was caught speaking into an open mic disparagingly about a voter he had just spoken to, calling her a ‘bigoted woman’. The story has dominated the news for the last two days.

In terms of the poll, a trend has been clear for the last week. The Conservatives have topped every poll for the last week, polling in the low to mid 30s. The Lib Dems have come second in most polls, closely followed by Labour. This is reflected in UK Polling Report’s polling average of CON 35, LD 28, LAB 27. Such a result would see the Conservatives win 282 seats, Labour win 250 and the Lib Dems win 86 under the uniform national swing model.

This model of calculating the effects of swings has been criticised by Nate Silver of US psephological blog FiveThirtyEight, who has produced a new sophisticated model. Silver’s model would see Labour lose far more seats, with the Conservatives winning a few more seats and the Lib Dems a lot more. Having said that, all of the projections produce a hung parliament with the Conservatives as the largest party, followed by Labour and the Lib Dems, who would gain a large number of extra seats without overtaking either major party.

Prominent UK psephobloggers Robert Ford and Anthony Wells have engaged in the debate, with Nate Silver responding, and they are well worth reading before trying to interpret the consequences of opinion polls over the next few days. While it is relatively simple to predict the general consequences of swings in elections with simple two-party races and small swings, but in such a dramatic race with massive shifts in polls, it is extremely difficult to predict the result across the constituencies.

I will be covering the election results on Friday morning Australian time as the UK results begin to flow in. I’m also expecting my UK election map to be used on the air for CNN International’s election coverage, so keep an eye out.

UK 2010 – Lib Dems surge

The UK held its first ever general election debate between party leaders on Thursday evening, and instant reaction polls showed a clear majority agreeing that Nick Clegg, leader of the Liberal Democrats, came out as the winner. Voting intention polls since Friday have indicated that the Liberal Democrats have experienced a remarkable increase in support, effectively making the race a three-way tie.

Four polls have been released today, showing that the vote between CON/LAB/LD is as follows:

  • YouGov/Times: 33/30/29
  • ComRes/Independent: 31/27/29
  • ICM/Telegraph: 34/29/27
  • BPIX/Mail: 31/28/32

While the order varies in all three polls, it is clear that the Liberal Democrats have jumped into contention with the other major parties. Every poll for the week before Thursday’s debate saw the Lib Dems in the 18-23% range.

It is not yet clear what effect this poll surge will have come election day. Clegg was clearly the unknown figure in the debate, and his strong performance and outsider positioning has been largely responsible for the increase in the Lib Dem vote. The two other parties have largely ignored Clegg for most of the campaign, but that has already begun to change. Will the enthusiasm from a single debate performance fade, or can the Lib Dems maintain their support by positioning themselves as a force capable of winning the election?

The Lib Dem surge isn’t solely due to a strong performance on a debate. It is rather the harnessing of years of disillusionment with the political system that reached a crescendo with the 2009 expenses crisis. Both major parties are held in very low esteem and Clegg’s message of change would have appealed to a great number of them. The Lib Dems have consistently polled in the high teens or low 20s for most of the last decade, and that strong third party force was well-placed to take advantage of such a crisis. They have the strength to be credible without being too tainted by the scandals and disillusionment.

Read the rest of this entry »

Tasmanian poll: Greens outpoll Labor

Another EMRS poll has been released over the weekend in Tasmania, which reinforces the trend seen in a February poll towards record levels of support for the Tasmanian Greens.

While it appears that the full breakdown of results has not been published online (only appearing in the Examiner), the topline figures are:

  • 29 – Liberal
  • 22 – Greens
  • 21 – Labor
  • 2 – Others
  • 26 – Undecided

All EMRS polls are published with high levels of undecided voters, meaning that these numbers would be much higher once those are taken into account, and undecided voters usually favour the major parties over the Greens, particularly that major party with the best shot of majority government, although on the current numbers that doesn’t seem likely at all.

The poll also broke down votes according to each electorate. Samples are usually too small to take them seriously, although much has been made of the Denison poll predicting that Scott Bacon, son of the former premier, would possibly take the only ALP seat, beating out Premier David Bartlett and ministers Lisa Singh and Graeme Sturges, showing the importance of name recognition in Tasmanian politics.

Since the last two polls were very similar, Tasmanian psephologist Kevin Bonham has combined the electorate breakdowns to produce a more solid sample. These figures suggest that the Greens are solidly on track to win 6 seats (one in every district and two in Denison) and even outpolling the ALP in Franklin, which could suggest Adam Burling would have an outside chance of winning a second Greens seat in Franklin, although he doesn’t have a high profile.

In other news, four former premiers (two Labor, two Liberal) have come out to warn Tasmanians against a minority government, although at this point it doesn’t seem clear how Tasmanians could vote to avoid a hung parliament. Peter Tucker has commented on this panic-stricken move at his Tasmanian Politics blog.

Update: William Bowe has managed to track down more information about this elusive poll over at Poll Bludger. After including “leaners” the figures come out as:

  • 30 – Liberal
  • 23 – Labor
  • 22 – Greens
  • “almost a quarter” undecided

Which he interprets as 39 Liberal, 30 Labor, 29 Greens. Compared to the February poll, this has the undecided vote up a large amount (from 13% to 23-4%), with Liberals down four, Labor down four, and the Greens down two.

Greens surge in Tasmanian poll

The Tasmanian campaign has kicked into gear today with a poll from pollsters EMRS showing a surge in support for the Tasmanian Greens. The poll has shown statewide results of 39% for the Liberal Party, 31% for the ALP and 27% for the Greens.

EMRS has also published breakdowns for the five districts, which shows the Greens polling 40% in Denison, although sample sizes of less than 200 per seat make these unreliable in measuring each seat’s vote. All the same, the statewide figures would suggest vote levels in each seat similar to those shown in the breakdown.

If these figures were translated at an election it would produce a result of 10 Liberals, 9 Labor and 6 Greens, with the Greens gaining a seat in Braddon and a second seat in Denison. A 2-2-1 split in favour of the ALP and Liberals would be produced in the four regional electorates, with the ALP losing two of their three seats to the other parties, producing a 2-2-1 split favouring the Liberals and Greens.

This is likely a high watermark for the Greens and it would be a surprise if they polled this strongly, but it suggests the Greens could be in with a shot in Denison, where they are running Deputy Lord Mayor of Hobart Helen Burnet alongside sitting MP Cassy O’Connor, although it seems unlikely that Burnet will make up for the loss of high-profile former MP and Greens leader Peg Putt, who retired in 2008.

Elsewhere: Pollbludger and Antony Green.

Death by a thousand polls

Well, three polls, at least.

In a moment of synchronicity, three of Australia’s four pollsters have released federal voting intention polls, all clearly showing a collapse in support for the Coalition and possibly spelling the end of Malcolm Turnbull’s leadership.

Read the rest of this entry »