Plain Politics

Where the parties stand

A sample-weighted trend across every published national poll, shown with the uncertainty attached. A single poll is a noisy measurement — the band matters more than the line.

Across 148 published national polls, with fieldwork to 16 August 2026, the trend has Labor on 28.0%, One Nation on 26.0%, Coalition on 21.9%, Greens on 13.0% and Others on 11.1% of the first-preference vote. The trend is estimated at 13 August 2026, the midpoint of the most recent fieldwork.

Labor
28.0%
95% range 27.0–29.0
One Nation
26.0%
95% range 25.1–27.0
Coalition
21.9%
95% range 21.0–22.8
Greens
13.0%
95% range 12.3–13.7
Others
11.1%
95% range 10.4–11.7

Primary vote trend

148 polls from 14 polling organisations, fieldwork between 5 May 2025 and 16 August 2026.

0%10%20%30%40%Jun 2025SepDecMar 2026JunLabor 28.0One Nation 26.0Coalition 21.9Greens 13.0Others 11.1
Shaded bands are 95% confidence intervals on a sample-weighted local trend (21-day bandwidth), widened by a factor of 1.7 to allow for house effects and other non-sampling error. Marks are individual polls. Where a pollster does not report a series it is shown as “not reported” and excluded — never estimated. Poll figures are facts published by the named polling organisations; where available each record links to that organisation's own release. Table structure and aggregation sourced from Wikipedia, 'Opinion polling for the next Australian federal election' (CC BY-SA 4.0).

Preferred prime minister

48 questions, as published

A different question from the one above: not which party a respondent would vote for, but which person they would rather have in the job. Pollsters ask it in different forms — some offer two names, some three — and a leader's figure moves with who else is on the list, so every question is shown as it was asked.

Most recent reading, by pollster

Different pollsters offer different people and give an undecided option different prominence, so their figures are not directly comparable with each other. Who each one has ahead is.

  • YouGov30 June 2026 · n=1,502
    Albanese 44Taylor 35unsure 21
  • RedBridge26 June 2026 · n=1,006
    Albanese 33Hanson 23Taylor 11unsure 23
  • Newspoll25 June 2026 · n=1,235
    Albanese 47Taylor 36
  • DemosAU18 June 2026 · n=1,497
    Albanese 35Hanson 28Taylor 19
  • Resolve13 June 2026 · n=1,801
    Hanson 33Albanese 29Taylor 16unsure 22
  • Fox & Hedgehog26 May 2026 · n=1,700
    Taylor 38Albanese 36unsure 26
  • Spectre Strategy8 April 2026 · n=1,002nothing newer from them
    Albanese 41Taylor 32unsure 27
  • Freshwater29 March 2026 · n=1,050nothing newer from them
    Albanese 42Taylor 36unsure 22
Albanese v Taylor30 polls · Albanese ahead in 28, Taylor ahead in 2
Fieldwork toPollsterAlbaneseTaylorUnsureSample
30 June 2026YouGov44%35%21%1,502
25 June 2026Newspoll47%36%not reported1,235
16 June 2026YouGov43%38%19%1,492
4 June 2026Newspoll44%38%18%1,240
2 June 2026YouGov41%39%20%1,471
26 May 2026Fox & Hedgehog36%38%26%1,700
19 May 2026YouGov41%38%21%1,504
17 May 2026Newspoll46%38%16%1,252
16 May 2026Resolve30%33%37%1,800
5 May 2026YouGov45%36%19%1,500
21 April 2026YouGov44%39%17%1,501
18 April 2026Resolve33%32%35%1,803
16 April 2026Newspoll46%37%17%1,235
8 April 2026Spectre Strategy41%32%27%1,002
7 April 2026YouGov44%36%20%1,502
29 March 2026Freshwater42%36%22%1,050
26 March 2026Newspoll44%36%20%1,232
25 March 2026Fox & Hedgehog39%35%26%1,810
14 March 2026Resolve35%31%not reported1,803
26 February 2026Newspoll45%37%18%1,237
19 February 2026Fox & Hedgehog40%35%25%1,625
14 February 2026Resolve38%22%40%1,800
10 February 2026YouGov47%25%27%1,561
8 February 2026Newspoll49%30%21%1,234
29 January 2026RedBridge37%9%54%1,003
27 January 2026YouGov47%29%24%1,500
16 January 2026Newspoll51%31%18%1,224
16 January 2026Resolve33%29%38%1,800
6 January 2026DemosAU42%29%29%1,027
6 January 2026Fox & Hedgehog39%31%30%1,608
Albanese v Hanson7 polls · Albanese ahead in 7
Fieldwork toPollsterAlbaneseHansonUnsureSample
30 June 2026YouGov49%40%11%1,502
16 June 2026YouGov48%41%11%1,492
2 June 2026YouGov47%41%12%1,471
19 May 2026YouGov50%38%12%1,504
5 May 2026YouGov54%35%11%1,500
21 April 2026YouGov50%39%11%1,501
8 April 2026Spectre Strategy44%39%17%1,002
Albanese v Hanson v Taylor11 polls · Albanese ahead in 10, Hanson ahead in 1
Fieldwork toPollsterAlbaneseHansonTaylorUnsureSample
26 June 2026RedBridge33%23%11%23%1,006
25 June 2026Newspoll49%31%20%not reported1,235
18 June 2026DemosAU35%28%19%not reported1,497
13 June 2026Resolve29%33%16%22%1,801
28 May 2026RedBridge31%25%14%30%1,005
20 May 2026DemosAU34%27%23%16%1,502
30 April 2026RedBridge33%23%14%29%1,014
27 March 2026RedBridge33%23%14%30%1,003
27 February 2026RedBridge34%23%10%33%1,006
20 February 2026DemosAU37%25%19%19%1,551
21 January 2026DemosAU39%26%16%19%1,933
Earlier leadership pairings20 questions · Albanese v Ley

Pollsters stop asking about a pairing once someone in it is no longer leader, so these series simply end. They are kept because they were really published, and held apart because a question about a former leader does not describe the contest now.

Albanese v Ley20 polls · Albanese ahead in 20
Fieldwork toPollsterAlbaneseLeyUnsureSample
20 December 2025Resolve38%30%32%1,010
12 December 2025RedBridge41%12%47%1,012
7 December 2025Resolve41%26%33%1,800
20 November 2025Newspoll54%27%19%1,245
13 November 2025RedBridge40%10%50%1,011
8 November 2025Resolve39%25%36%1,804
30 October 2025Newspoll54%27%19%1,265
30 October 2025YouGov51%26%23%4,578
20 October 2025Freshwater48%31%22%1,530
12 October 2025Resolve40%23%37%1,800
2 October 2025Newspoll52%30%18%1,264
30 September 2025YouGov50%28%22%1,329
13 September 2025Resolve38%26%36%1,800
11 September 2025Newspoll51%31%18%1,283
16 August 2025Resolve41%26%33%1,800
14 August 2025Newspoll51%31%18%1,264
30 July 2025Wolf & Smith45%35%20%5,000
18 July 2025Resolve40%25%35%2,311
17 July 2025Newspoll52%32%16%1,264
1 July 2025Spectre Strategy46%27%27%1,001

How to read these

  • A leader’s score depends on who else was offered. The same leader in the same month scores differently in a two-way question and a three-way one, so those are listed separately and never averaged. There is no combined line here because no pollster published one.
  • Pollsters differ in how prominently they offer an undecided option, and the more prominent it is the lower every leader scores. Compare the gap between leaders within a poll, not a leader’s number across pollsters.
  • This is not voting intention, and it is not a seat projection. It asks which person a respondent prefers, which is a different question from which party they would vote for.

Figures are facts published by the named polling organisations; where available each record links to that organisation's own release. Table structure and aggregation sourced from Wikipedia, 'Leadership approval opinion polling for the next Australian federal election' (CC BY-SA 4.0).

What that would mean in seats

A model, not a measurement

A national vote is not a seat count. Seats are won one division at a time, so a party polling well everywhere can win far fewer seats than its vote suggests, and a party whose vote is concentrated can win far more. This takes the 10 most recent polls (2026-07-26 to 2026-08-16), applies them to each division’s 2025 first preferences, and distributes preferences as they actually flowed in 2025.

How a national swing lands in each seatSwing
How Coalition voters preference One NationCoalition preferences
76 for a majority
  • Labor93(59–111)
  • Coalition48(17–63)
  • Others7(1–14)
  • One Nation1(0–59)
  • Greens1(0–4)

Bold is this projection. The range in brackets is what the same model gives across 117 combinations of swing method and of each party polling at the top and bottom of its recent range.

NSW
46
VIC
38
QLD
30
WA
16
SA
10
TAS
5
ACT
3
NT
2
One square per seat, grouped by state and sorted by projected winner. A dot marks a seat changing hands — 19 of 150. Hover or tab for the detail.

What is measured and what is assumed

  • Measured:every party’s first preference vote in all 150 divisions in 2025, where preferences went when a candidate was excluded, and the current national primary vote from published polls.
  • Assumed: how a national change lands in an individual seat. A party rising from 6.4% to 26.1% could add the same number of points everywhere, or multiply wherever it already stood. Those give different answers, so both are offered and the gap between them is published as the range.
  • Where that breaks: multiplying works for a party moving a few points and fails for one moving twenty. Because One Nation’s national vote quadruples, the proportional setting puts them above half the first preferences in ten seats and above 40% in twenty-five — when the highest first-preference share any party took in any seat in 2025 was 55.2%, Labor in Sydney. Read the top of the range as a bound the method produces, not as a forecast.
  • Assumed: how Coalition voters preference One Nation. This is the weakest number in the model and the one the answer turns on. The 2025 figure is 5.2%, but it comes from the only 25 divisions where a Coalition candidate was ever excluded — Melbourne, Sydney, Grayndler, Canberra and the like, safe Labor and Green seats where the Liberals ran third and their preferences went to Labor. Nothing in that sample resembles a regional contest between One Nation and Labor, which is where this projection now puts them. Set it to mirror how One Nation voters preference the Coalition (61%) and One Nation goes from a handful of seats to dozens. Both are on the control above; neither is a fact.
  • Checked: handed the 2025 vote, the model returns 143 of 150 divisions correctly. The 7 it misses are seats decided by a candidate rather than a party — Bullwinkel, Curtin, Fowler, Kennedy and others — which is the blind spot of any model built on national figures.
  • Independents are modelled apart from minor parties, because an independent’s vote is personal to one seat while minor party support is spread thin across all of them. Lumped together, a national fall in the minor party vote is subtracted from a sitting independent’s own vote and takes their seat off them. They are still counted as one crossbench, which held 12 of 150 seats in 2025. Independents are held at their 2025 share, so where a sitting independent is not recontesting this will be wrong.
  • Seats are grouped by state and nothing finer. The data records which state a division is in and no more, so there is no map here — a shape claiming to be Australia would be asserting geography this model does not have.
  • Boundaries are those of the 2025 election. A redistribution since is not reflected, because sitting members were elected on these boundaries.

Seat baseline and preference flows from the Australian Electoral Commission (CC BY 4.0). Polls as listed above.

Known limitations of this dataset

8 of 148 records carry a caveat. We publish these rather than quietly dropping or patching them, because a silently “clean” dataset is the easiest kind to mislead with.

  • Essential · fieldwork to 27 Jul

    Published two-party-preferred (48/47) does not sum to 100 — typically an undecided share is included. Excluded as not comparable.

  • Essential · fieldwork to 29 Jun

    Published two-party-preferred (46/48) does not sum to 100 — typically an undecided share is included. Excluded as not comparable.

  • Essential · fieldwork to 25 May

    Published two-party-preferred (48/47) does not sum to 100 — typically an undecided share is included. Excluded as not comparable.

  • Essential · fieldwork to 27 Apr

    Published two-party-preferred (47/49) does not sum to 100 — typically an undecided share is included. Excluded as not comparable.

  • Essential · fieldwork to 23 Mar

    Published two-party-preferred (46/47) does not sum to 100 — typically an undecided share is included. Excluded as not comparable.

  • Essential · fieldwork to 24 Nov

    Reported primary votes sum to 102.0%, not 100% — transcribed as published.

  • YouGov · fieldwork to 30 Sep

    Reported primary votes sum to 107.0%, not 100% — transcribed as published.

  • Wolf & Smith · fieldwork to 30 Jul

    Not reported by this pollster: grn, onp, oth. Recorded as null, never estimated.