There Must Be Some Way Out Of Here: YouGov and DemosAU Tasmanian Polls

YouGov Liberal 31 Labor 34 Green 13 IND 18 other 4

DemosAU Liberal 34 Labor 26.3 Green 15.1 IND 19.3 other 5.3

IND vote likely overstated in both polls

Seat estimate if YouGov poll close to accurate 13-14-4-4 (Lib-ALP-Grn-IND)

Seat estimate for DemosAU 13-11-5-4, 2 unclear 

This article is part of my 2025 Tasmanian election coverage.  Link to main guide page containing link to other articles including electorate guides.  

At the 2024 Tasmanian election, voters elected a parliament where it wasn't easy to form a government at all, and the one that was formed didn't last for long.  Labor was unwilling to even try to form a government that would have involved the dreaded Greens, and the Liberals were only willing to form a government with what was left if it was basically a Liberal government with relatively minor concessions to others.  When that ceased to be a viable option upon the loss of the key vote of Andrew Jenner, the government was unable or unwilling to adjust to the fact that it was hanging by Craig Garland's fishing line, and here we are.

A new election offers the prospect that someone might break through and we might have some sort of a normal government - if not a majority (which needs a very hefty swing) then at least a stable minority government needing the support of two or three crossbenchers and able to find such numbers that they can work with.  But also, the quagmire might continue.  If neither major party appeals then we could even end up with a parliament where neither major party can govern without the Greens.  What happens if Labor wins the most seats but needs the Greens or needs, say, all of five other crossbenchers three of whom may as well be Greens?  What happens if we end up more or less back where we were?  (Only one state government in Australian history has ever pulled that off after losing a no-confidence vote).  

Tonight we got some public results from two polls, from YouGov and DemosAU.  The YouGov one is the more positive for Labor (the DemosAU one is horrible for them in the circustances) but at this stage neither shows either major party close to getting us out of the mess.  

YouGov

This poll was taken online from 12-24 June with a sample size of 1287 (Bass 253, Braddon 250, Clark 251, Franklin 266, Lyons 267).  Respondents were offered the choice of Liberal, Labor, Greens, Independent, or an Other (specify) option which I hear drew some predictably unavailable responses.  

These are the results from the YouGov website: 

The individual electorate samples of c. 250 shouldn't be taken very seriously.  Aside from them having a notional margin of error of over 6%, the real error margin is likely to be much higher because of weighting, targeting and sample pool effects (as with all single seat polling).  Nobody should believe the Greens with Rosalie Woodruff on top of the ticket are on only 9% in Franklin after polling 10.5% in the federal election with an ineligible candidate who had withdrawn from campaigning.   30% independent in Clark is also a major stretch.

At the 2024 election pollsters generally overestimated the independent vote despite the number and diversity of indies on offer.  It's not easy to put a number on this because some polls lumped independents and others or did other unusual things, and because full details of two media-reported Freshwater polls were frustratingly never obtained.  Polls with independent broken out averaged nearly 14% but independents only actually got 9.6%.  This is a problem with offering independent as a generic option - some voters think that Andrew Wilkie will be running in their state electorate and some also seem to confuse minor parties and independents.  There could well be a swing to independents this election with JLN not running, at least one new independent who could poll heftily (Peter George) and so on but 18% seems unlikely.  

Taking the poll numbers literally and assuming a fair degree of scatter in the independent vote, Bass would be 3 Liberal 2 Labor 1 Green 1 Independent (or 3-3-1-0), Braddon probably 3-3-0-1 though with some potential for a second independent, Clark 2-2-1-2, Franklin 2-3-1-1, Lyons 3-3-1-0.  That would be something like 13-13-4-5 (the scenario where either party needs the whole crossbench to get around the Greens) though there's a fair chance that Labor's 3% statewide primary vote lead would in reality be good for an extra seat somewhere.  Factor in the independent vote being likely to be overestimated and something like 13-14-4-4 looks like a better read.   I don't personally think the Greens are going to lose their second Clark seat to an independent anything like as easily as in this sample (if at all, which is not to say they won't lose it to Labor) but the numbers are the numbers and it's difficult to read this poll as "saying" anything different.

Other stuff in this poll includes a 43-36 lead for Jeremy Rockliff over Dean Winter as Preferred Premier.  Of course, I always prefer to see approval rating polling.  Preferred Premier polling skews to incumbents but this is at least hardly emphatic rejection territory for Rockliff.  The poll includes issue findings where voters are asked to choose between issues of health, debt, privatisation, the stadium and salmon farming.  Health comes out on top among these issues, especially in the north, and while voters tend to be against the stadium and privatisation, in neither case is this overwhelming.  Accounting especially for the cancelling out of for and against views, the stadium is significant, but its salience can be overestimated.  The government is running very hard on health announcements.  

DemosAU

I hope we see a lot more results from this poll, which was widely reported in field by people taking it via the EMRS survey portal.  At the moment what we have is a report in The Advocate regarding a poll taken for an "unnamed peak body" by a mix of robocall and panel methods and a very large sample size (4289) between June 19 and 26.  The poll has results of Liberal 34 Labor 26.3 Green 15.1 and "independents" 19.3, leaving 5.3 for others.  The Advocate has published the Braddon figures which are 44-25.2-9.3-15.6 leaving 5.9 for others.  The Advocate has interpreted these numbers as either 3-2-1-1 or perhaps 3-2-0-2 but a 15.6% independent vote in Braddon would most likely scatter between a number of candidates with only one (presumably Craig Garland) competitive.  Actually given their potential ability to spread votes between Gavin Pearce, Felix Ellis and Roger Jaensch, these numbers are not far short of four Liberals at the expense of the Greens or maybe Garland.  This makes sense given the Liberals almost got four last time with JLN taking a seat.

Neither poll shows anything promising for Tasmanian Nationals, who were named in the DemosAU poll.

I have seen some of the text of this poll and it tried to avoid the problem with generic independents by naming specific independents, however it did include some that are not running (eg Ben Lohberger was available as an option in Clark).  Even when knowing the candidates, naming the independents can also overestimate their vote in Hare-Clark because it means they are named while the major party candidates are not - this is one of the reasons Hare-Clark is so difficult to poll!

While this poll may be dismissable if there's a suggestion that the "peak body" was somebody adverse to Labor's interests, overall Labor would not want to see polls like this being talked about at all!  Polls that show them going backwards or even not going anywhere play into the government's narrative that Labor caused the election without knowing what they are doing and are still not ready to govern.  The YouGov poll on the other hand is more positive for them.

It's very difficult to interpret the DemosAU numbers in seat tally terms because of the very high independent vote, and for now without full breakdowns, but if Labor are going backwards on primary votes then even after accounting for the Independent overestimate they're probably not gaining many or perhaps any seats.   For the majors it seems something like the status quo.  Let's make it the status quo and say for now 14-10-5-6, with reservations about whether the 6 are independents, something else or maybe some don't even exist.   

I should add a cautionary note about this DemosAU in particular.  The EMRS online panel is an opt-in with a high percentage of political polls and a very high proportion of politics junkies are on it.  I suspect the EMRS online panel has the sort of skews towards high engagement voters that no degree of weighting will ever fix.  DemosAU polls in the federal election underestimated the combined major party vote, albeit in the case of their final poll only by 2.4% which isn't massive.  A degree of caution hence about the combined major party vote being quite as low as 61.3, especially with JLN not running and only three parties on the ballot in two electorates.  

I am aware of a third (private) poll which I may say more about that falls somewhere between these two, with a far lower but still quite high independent vote (around 12%) and the Liberals slightly ahead, with both majors in the low 30s.  Nobody anywhere near a majority in any of these, indeed no poll yet with a major party over 35.  [UPDATE 10/7: This poll was by uComms 10-11 June, with primaries after redistributing undecided of Liberal 33.7 Labor 32.7 Green 13.8 IND 11.9 JLN (not running) 4.0 SFF 3.3 other 0.9]

At some stage before the election I will again do the best I can to aggregate all these polls and others still to come.  I think it's a bit too early right now with the potential for more detail to emerge re DemosAU especially.

Tuesday: Some more detail on DemosAU; the Nats are on 2.3% statewide including 2.6% in Braddon and 3.7% in Lyons (that suggests they might be around 5% in Bass if they were only offered where running).  They trail SF+F both in Braddon and statewide.  In what are reported as if forced choices (but I am not sure about that yet and I would not recommend forced choice for the second question) voters are reported as preferring Rockliff to Winter 56-44 and being more likely to blame Winter than Rockliff for the early election 55-45.  

Wednesday: Pulse has published the full seat results with graphics and further details.  My take on Braddon is above.  For the remainder:

* Bass (33.5-27.5-18.8-IND 11.1-Nat 5.1-SFF 4.0): the Greens number looks too high, but Labor has not much over two quotas here.  A murky reading in terms of the potential for nobody much to have enough for a second seat; on these numbers 3-2-1 and the last one is possibly Pentland or a National, but Labor could be in the hunt if their non-Finlay vote split evenly between two candidates.  (Two Greens couldn't even be ruled out on those numbers but adverse preference flows would probably take care of that).  

* Clark (26.2-23.6-22.7-27.5) Although the Independent vote combined is over two quotas it is going to be overestimated and will include a lot of scatter as different indies get excluded, this is clearly 2-2-2-1

* Franklin (29.1-22.6-12.9-35.4) The Independent vote here is enormous and I can't take that seriously but anything remotely near that is going to be 2-2-1-2 with both O'Byrne and George elected.  The Liberals are too far short of three for even a perfect split between their candidates to save them as they will get killed by preferences.  

* Lyons (35.9-31.9-13.1-IND 8.4-SFF 7-Nat 3.7) this is 3-2-1 and of all things Labor vs SF+F for the final seat though the SF+F vote looks excessive; they only got 4.8% with 5 candidates in 2022 so let's say 3-3-1

My estimate of the total then for this poll is 13-11-5-4 (Liberal-ALP-Green-IND) with two unknown; Labor would not get to 18 without the Greens; the Liberals would need the whole or nearly the whole non-Green crossbench.  The poll also finds 61% wanting a majority government - which is about as many as it finds intending to vote for both major parties!

Can also (again from DemosAU via Pulse) add the excluded undecided votes: Bass 9.5 Braddon 12.9 Clark 9.8 Franklin 6.3 Lyons 13.7.  

Thursday 3 July: Something I should note re the very high Clark IND vote in both polls is that votes are not readily transferrable between different independents - there is some flow but it's not necessarily that strong.  If INDs were to get more than two quotas in Clark that could well put Kristie Johnston way over a quota.  However in that case her surplus would not necessarily flow much to Elise Archer (though they are adjacent on the ballot paper which helps).  This is all of course speculation anyway - Johnston got 0.61 quotas last time, down from 0.88 in current terms in 2021.