South Australia, October 2025

01

The two-party figure is struck against a party that would finish third

The lower-house question offered four options. The upper-house question listed fourteen and put One Nation on 12 per cent. At the election five months later One Nation took 22.9 per cent of the lower-house vote and the Liberals 18.9.

The lower-house figures exclude 7 per cent of voters who were undecided and the upper-house figures 12 per cent. Both columns sum to 100. The ten further upper-house entries run from the Animal Justice Party, Legalise Cannabis and SA-Best at 4 per cent each down to one group recorded above zero but below 0.5.

Labor and the Liberals each lose ground when the longer list is offered, ten points and four points, and the parties absent from the shorter list take 35 points of the upper-house vote between them. The pattern is the one Victoria, New South Wales and Western Australia all show in the same months: a four-option lower-house ballot concentrates support on the parties it names and hides the rest in a category.

South Australia is where that mattered most. One Nation's 12 per cent here was its upper-house figure in October; by the February wave it was on the lower-house ballot at 19 per cent, and at the 21 March election it took 22.9 per cent and finished ahead of the Liberals on 18.9. The 66–34 on this page is therefore a two-party preferred against the party that would come third, calculated five months before an election that produced a One Nation opposition. It is an accurate record of what this question measured and a poor guide to what was coming, and both halves of that are worth carrying forward. ACE-IU-2026-034 reads the wave where the question changed.

02

The Liberals do not reach a third of any cohort in the poll

DemosAU publishes no cohort sample sizes for this wave. If the three-way splits each hold about a third of the effective sample of 661, a two-party gap between two of them carries roughly ±9.3 points.

Every first-preference row sums to 100 in the report. The undecided column sits outside that total and is excluded from the party shares. Cohort figures rest on subsamples and carry wider margins than the ±3.9% statewide topline.

The Liberal vote runs from 13 per cent among renters and voters under 35 to 32 among outright owners. It clears 30 in two cohorts of fifteen and reaches a third of none of them. Labor's floor across the same table is 41. A party in opposition whose best segment of the electorate gives it under a third, five months out from an election, is not in a position to be recovered by a campaign, and it was not: the Liberals polled 18.9 per cent in March.

Two divisions clear their intervals. Labor's two-party share runs from 73 per cent among under-35s to 59 among the over-55s, a fourteen-point gap, and from 73 among renters to 55 among outright owners, an eighteen-point gap. The Greens' 29 per cent of under-35s against 6 among the over-55s is a 23-point gap and the widest spread any party shows across the cohorts here. The eight points between men and women on the Labor primary sit at the edge of an interval of about ±7.6, so that one is a reading to watch rather than a finding; the eleven points between them on the Greens vote clear about ±5.1 and are a real difference.

03

Ramping and the algal bloom are rated as badly as cost of living

Asked whether the Malinauskas government is doing a good job on each of ten issues, respondents put it in positive territory on three. The two state-specific issues sit at the bottom with cost of living.

Each bar is the net rating on an issue, the good-job share less the bad-job share, in points. The note gives the three underlying shares.

Hospital ramping is the issue this battery exists to measure and it is rated 42 points negative, level with cost of living. A government carrying that rating on the issue most identified with it, while leading 66–34, is the clearest illustration in this series that issue ratings and voting intention are separate measures. South Australians rated this government badly on most of what it does and intended to re-elect it comfortably, and in March they did.

The algal bloom at 38 points negative is the state-specific entry and it is not a minor one: 6 per cent of respondents name it unprompted as the biggest issue facing South Australia, ahead of crime at 5 and climate change at 3. The Whyalla steelworks runs the other way, at 6 points positive on a 58 per cent neutral share, which is what an issue looks like when most of the electorate has no settled view and the rest leans favourable.

04

October against the 2022 election

Percentage points, lower-house first preferences. Labor is up seven on the result that brought it to office and the Liberals are down almost fifteen.

The changes are DemosAU's own. The any-other-candidate baseline of 15.2 is the 2022 others result of 12.6 plus One Nation's 2.6, because this question had no separate One Nation box; the February wave, which does, measures others against the 12.6 alone. The two-party baseline of 54.6 is the figure DemosAU's stated change of +11.4 implies.

Three and a half years into the term the government's primary vote was seven points above its election result and the opposition's was almost fifteen below it. Nothing in this wave suggested a contest. What it did not and could not show was the shape of the alternative: the question offered one, and by March the electorate had chosen a different one. Read beside ACE-IU-2026-014, which scores the final pre-election poll against the result, this wave is the record of the last moment at which South Australian politics still looked like a two-party contest.