Dutton’s Sliding Doors: What Polling Tells Us About the 2025 Election

During the 2025 federal election campaign, the Liberal leader Peter Dutton referred to the election as a ‘sliding doors’ moment for Australia several times. He presented the poll as a stark choice between two visions for Australia.

Dutton’s comment invites an inquiry not just into how the election turned out, but also how it might have turned out. In the movie “Sliding Doors”, the life of the lead character played by Gwyneth Paltrow is divided into two parallel universes, determined by one trivial decision. Peter Dutton’s and the Liberal Party’s mistakes during the 2025 election campaign were not, however, the result of one, or even a series of, trivial decisions – they were the product of a cavalcade of misjudgements.

No doubt Dutton meant to echo Paul Keating’s comment made just before the 1996 election, ‘If you change the government, you change the country’. Keating’s effort to stop people from voting for John Howard didn’t work. Peter Dutton’s ‘sliding doors’ effort was also unsuccessful. In 2025, Australian voters didn’t accept the picture of an existential crisis that Dutton presented. Voters couldn’t see anything catastrophic about either choice and settled for what they knew. While 94 seats gives Labor and the Prime Minister a sense of dominance, it shouldn’t. Voters were not particularly committed to the decision they made. They voted Labor less because of their faith in the ALP, and more because of their doubts about the Coalition.

The Polling

For 24 years I have been conducting online qualitative research to understand what motivates voters. Over the course of 2025, I conducted three surveys to gauge the public mood. There were 712 respondents in the first survey, conducted in January; 762 respondents in the second survey, conducted at the end of March and early April; and 626 respondents in the final survey, held over the weekend of the election at the beginning of May. As with any qualitative research, care must be taken with the data, as it is indicative rather than representative. However, even so, the quantity of data allows for both quantitative results and qualitative feedback.

Some respondents have been participating in my surveys and providing me with information about their voting habits and motivations since I first began conducting polls twenty-four years ago. Others have been recruited over the years since. My group of respondents skews left, is older and more male than the population, and requires techniques to normalise the response. One of the things that is most useful about this group is that these are people interested enough in politics to give me fifteen minutes of their time per survey to share their thoughts and opinions. They’re better informed than many voters, but they’re not experts; they’re simply more interested in politics. The benefit of this is that they can express their thoughts and feelings more effectively than the average person, and they tend to reach similar conclusions, albeit earlier in the process, making them ideal for election analysis. It also means that they have more settled opinions and often cluster closer to the extremes than a randomly selected group of respondents would. When the results of my qualitative research are combined with publicly available quantitative polling, we can piece together the elements of the drama of the last election and then examine how the election unfolded.

In January 2025, Peter Dutton had cause for optimism. Roy Morgan set the scene at the start of January with the Coalition on 40.5% first preferences and Labor on 31%, giving the Coalition a two-party preferred vote of 53% to 47% – a clear majority. This finding was broadly consistent with polls from YouGov, Resolve, Newspoll, and the Coalition’s own pollster, Freshwater Strategies. The only outlier is the trade union-backed pollster Essential, which nevertheless still had the Coalition with a narrow lead. Labor, with good reason, would have had cause for concern. I know that Labor was told by its own pollster, Redbridge, that the ALP was behind the Coalition, and that Labor’s base vote was starting to fracture, and that around half of all voters were ‘soft’ and ‘in-play’. People thought the country was on the wrong track (rhymes with Dutton’s election slogan), and that this was driven by cost-of-living pressures, in particular housing. It was an environment where electors didn’t feel either main party represented their values, and that there was a growing cohort of ‘Dutton-curious’ voters.

Redbridge pinpointed outer suburbs and regional cities as the heart of voter discontent. Redbridge had nine seats changing hands, along with 14 seats too close to call. There was mention that migrant communities were also starting to move towards the Coalition.

It’s difficult to know how much of this was reflected in the Coalition’s research, but there’s no reason to think that the Coalition’s polling differed from the public polls. What we do know, though, is that as the opposition grew increasingly confident, hubris developed, and the Coalition, inexplicably, failed to manage expectations. Compulsory voting married to the Australian character means you always want to be the underdog going into an election, hence the tried and trite politician’s phrase, ‘I don’t pay attention to the polls, the only poll that counts is the one on the day.’

Australia is a country that likes a handicap and takes a Victa lawn mower to tall poppies on the weekend. If you’re the opposition you want all the electorate’s attention on the government because ‘governments lose elections, oppositions don’t win them.’ When voters think you might win, they start looking more critically at your policies, and may decide to vote you out before they have voted you in. Sympathy for your opponent, or some other extraneous feeling, can also cause votes to stray, particularly when maybe half of the vote is soft.

This is a plot element – that the Coalition was a strong possibility to win the election – the Coalition should have tried to modify, but it didn’t. In fact, it seemed to embrace it, with Peter Dutton even musing later in the campaign about which great house – Kirribilli or the Lodge – he would live in when he became prime minister.

I believe there was another consequence of the positive polling for the Coalition. I think the opposition decided conditions were similar to those in the Queensland election which saw David Crisafulli convincingly oust Labor with a very focused and ‘small target’ campaign. As a Queenslander, Peter Dutton would have seen firsthand the Liberal National Party’s campaign for the state election in October 2024, which saw the LNP gain a two-party swing from the ALP of 7% and win an additional seventeen seats. What happened at the Queensland state election could explain how Dutton navigated the next three months, as though the election was his to lose, rather than his to help the government lose.

The Coalition should have been sceptical of the predictive power of the polling. In Australia, it is more the exception than the rule for oppositions to come from behind and win an election. In January 2025, Peter Dutton and the Coalition had been ahead in the public polling for a bare three months. Compare that to Kevin ‘07’ Rudd, who had been polling in the mid-50s for the best part of a year, or to Tony Abbott who before his election victory in 2013 had been ascendant in the polls for over two years. The polling for the Coalition in late 2024 and early 2025 was an invitation to win the election, not the confirmation of success.

Then there were the issues. Clearly cost-of-living was a significant issue, but there were also subsidiary ones. And issues have to be triangulated between their salience, whether they will actually change votes, and which party is regarded as best to deal with them.

We have two glimpses into what electors were thinking. The first is courtesy of the IPSOS Issues Monitor in April 2025 which gave polling results up to March. It showed that cost-of-living was a major issue for 64% of voters, followed by housing (37%), healthcare (29%), the economy (28%), and crime (24%). Other issues of concern to voters were immigration (15%), the environment (11%), and defence (9%). Of the five most important issues: cost-of-living, housing, healthcare, the economy, and crime – there was only one issue, crime, on which the Coalition was judged to be more capable of managing than the Labor Party (29% of respondents said the Coalition was the most capable of controlling crime compared to 26% for Labor). Notably, 32% of respondents believed the ALP was the most capable of managing the economy, compared to 30% who said the same of the Coalition. There was no issue on which either side had established a dominance – except for health. 35% of respondents believed Labor was better at managing health, compared to 24% for the Coalition. Overall, the message to the major parties from this polling should have been sobering. On most issues, up to a third of voters were undecided or didn’t know which party would be better at managing the issues facing the country.

My survey also provides data on the perceptions of the parties. It doesn’t completely align with Ipsos, and we didn’t measure for exactly the same things. We asked respondents to rank a series of issues in order of importance. As Ipsos did, we chose the issues, but it was a much shorter list and they were issues that I thought could be decisive and which either of the two majors might run on or be unable to avoid.

There were seven measures: cost-of-living, housing affordability, taxation and the economy, healthcare and social services, climate change and the environment, national security, and immigration. I find the most useful way of measuring the strength of the rankings is to sum the percentage of respondents who nominated each issue as one of their top three issues. On this basis cost-of-living was ranked as the most critical issue (31%) followed by health (28%), tax and the economy (26%), housing (23%), climate change and the environment (18%), and finally, national security (15%).

I also asked which party the respondent believed was best to manage these issues. On cost-of-living the Coalition was preferred by 42% to 38%; on health Labor was preferred 47% to 33%; on housing the Coalition was ahead 36% compared to 32%, and on economic management the Coalition was favoured 43% to 40%. A perhaps surprising result was that the Coalition was preferred on climate change and the environment, 37% to 30%, and on national security, it was also preferred, 49% to 33%.

So what we can say is that the parties were fairly evenly matched on all the crucial issues, but that Labor had a strong lead on health, and the Coalition a strong lead on national security. However, a significantly higher proportion of respondents regarded health as a more important issue than national security. Hence, the leverage afforded by Labor’s strongest suit was much greater than that afforded the Coalition by theirs.

The mood of electors was ugly. Respondents believed the country was heading in the wrong direction. To examine I split my sample into five ‘tribes’: Labor, Coalition, Greens, ‘Nationalists’, and ‘Cosmopolitans’. The first three are self-explanatory.

‘Nationalists’ align with conservative, libertarian, and right-wing populist minor parties. They believe that a country has a soul, rather than just being lines on a map; that you don’t spend more than you earn, and you don’t borrow what you can’t repay; and that family, if not God, is important.

‘Cosmopolitans’ describes many of the voters who vote for the Teals, as well as the variety of left-wing populists and socialists. ‘Cosmopolitans’ think of Australia as comprising anyone who happens to live here; they have an entitlement mentality believing that the announced needs of any person or group in society who claim to be marginalised must be met by the resources of the country, by redistribution if necessary; and they adhere to varying degrees of ‘wokeness’.

Only 30% of my respondents thought the country was heading in the right direction, compared to 56% who thought it was not. Only ALP voters believed the country was heading in the right direction, while 93% of Coalition voters and 92% of ‘Nationalists’ thought it wasn’t. 41% of Greens voters thought the country was heading in the wrong direction compared to the 29% the country was heading in the right direction, while of the ‘Cosmopolitans’ 53% thought the country’s direction was wrong, compared to 21% who thought the country was heading in the right direction. The gripe of Greens and the ‘Cosmopolitans’ was generally that Labor was not going far enough on issue like climate change and income redistribution. These findings suggest that voters were grumpy, but reveal less about the direction the country’s voters would prefer.

On their own account, voters had good grounds to be grumpy. They didn’t think either major party leader was doing a good job. Only 30% approved of Anthony Albanese’s performance, while 58% disapproved. Dutton fared slightly better. Thirty-eight per cent approved of him, and 52% disapproved.

When asked whether the government deserved to be re-elected, only 39% agreed, with 52% disagreeing. Only 37% believed the opposition deserved to win the election, and 53% said they didn’t. Voters had a choice between a government they were sure couldn’t govern, and an opposition that voters couldn’t be sure about either. The upshot of all of this was that in mid-January, 25% of voters thought the ALP would win the election, 36% thought the Coalition would, and 40% expected a hung parliament. But what voters thought would happen wasn’t necessarily their desired result. When asked what they wanted to happen, 47% nominated a Coalition win, with 39% for Labor, and 14% hoped for a hung parliament. When it came to voters’ preferred prime minister it was a tie. Each of Albanese and Dutton was preferred by 47%, with 6% undecided.

In January 2025, neither camp could have been confident of victory. While there was a big primary vote gap between the two parties with the Coalition on 39% and Labor on 33%, the conventional wisdom is that either party needs a primary vote about 40% to be confident of victory. The problem with this traditional wisdom is that 2022 proved it wrong, when the ALP won with a primary vote of 32.5%, three percentage points lower than the Coalition.

This conventional wisdom had been shown not to apply to the ALP in 2022 when Albanese won with a vote of 32.6%, three percentage points lower than the Liberals on 35.7%. Labor was greatly assisted by voters for the minor left-wing parties being more likely to preference the ALP, then voters for minor right-wing parties are likely to preference the Coalition.

Neither party had a charismatic leader and as discussed on the issues Labor had a slight advantage. Health, the issue where they were dominant, is a concern not just of Labor voters but of all the other voter tribes so it had potential to increase their support. The Coalition had national security, but the only tribes concerned about that were Coalition and Nationalist voters, so that advantage was only likely to intensify the Coalition’s existing support, and potentially winning it preferences, but not much else.

By 28 March when Albanese announced the date of the election as 3 May, the campaign had been underway for some months and Labor were starting to gain the upper hand against the Coalition. Our sample indicated that Dutton’s approval rating was declining, while Albanese’s was slowly improving, although both still had a net negative approval rating. Increasingly ‘undecideds’ were favouring Labor, and expectations of a hung parliament were growing. However, what voters wanted to happen was virtually unchanged – their first preference was a Coalition victory, followed by a Labor victory, followed by a hung parliament. Voters found the Coalition’s slogan, ‘Getting Australia back on track’ more convincing than Albanese’s ‘Vote Labor so we can keep building Australia’s future together.’

By the weekend of the election, our survey indicated a decisive shift to the ALP. Albanese almost had a net approval rating (he was minus 1%), while Dutton’s approval had plunged (he was minus 42%). Voters thought that Labor had done enough to be re-elected.

As we now know the actual Labor two-party preferred vote was 55% gained with a with a primary of around 35%, a figure that had barely moved since the beginning of the election campaign. By contrast the Coalition first preference vote had slumped by around seven points to 32%.

Explanations of What Happened

Several factors contributed to the Coalition’s defeat at the 2025 federal election. As mentioned, the Coalition was arrogant and failed to impress the electorate.

There were undoubtedly polling failures. We know that the Coalition’s pollster thought they had a path to victory right up until the final weeks, because he said so publicly. But it is impossible to see where that path was.

However, all the pollsters were shockingly wrong in their predictions. In the last week the median estimate of the two-party preferred vote for the ALP was 52.2% and all the polls were tightly clustered around that with a standard deviation of only 0.00779. This means the actual result was almost four standard deviations away, and the chances of this are about 1 in 10,000.

This suggests either a common assumption that does not accurately represent the real world, or clustering – a process where pollsters adjust their figures to be close to where they think the other pollsters are. This is a well-known reputation risk minimisation strategy that is not helpful in good political analysis.

However, the bad polling should have had little effect on the campaigns. Polling in campaigns is useful for two main reasons: to determine where to allocate resources and whether your messages are resonating.

I’ve picked at random a few seats that changed hands at the election and analysed what occurred and there doesn’t seem to be any sign that resources as measured by money made a difference.

For instance, we know that Peter Dutton spent a significant amount of money in Dickson and lost, with a margin of -7%. Next door in the seat of Petrie, where the incumbent, Luke Howarth, thought he was safe and had donated money to other seats, he would have spent less than Dutton. In Petrie the swing against the Coalition was about 6%. Measuring the impact of messages is not an exact science, but if your messages are having an effect, even polling in the wrong seats should still show some impact from your advertising. Above anything, the polling could not tell us what was going to happen.

Three issues have been raised as potentially fatal to Dutton’s campaign.

First there was the Coalition’s nuclear energy policy. The claim made by Labor, the Greens, and a group calling themselves ‘Liberals Against Nuclear’ is that the policy cost the Coalition votes. It’s certainly true that at least initially Dutton was identified more closely with his ambition for nuclear energy, than with any other policy. If you wanted a sign of vision and bravery, this was it. And for a long time, there were no other significant picture issues to compete. However, by the time of the election our respondents were telling us that nuclear was very much a secondary issue. Those who raised it as a negative were most likely either people who would never vote for the Coalition, or Coalition or Nationalist voters who approved of the policy.

From our polling there’s little evidence that Labor’s ‘Mediscare’ campaign, in which the ALP claimed the Coalition would cut health funding to pay for nuclear power stations, had much traction either. It didn’t show up in our conversations. Inasmuch as it had an effect I suspect it was because it drew people’s attention to Labor’s $8 billion additional expenditure on health, which emphasised a policy area where Labor had dominance and salience.

There was also a beneficial side effect of the nuclear policy – climate change ceased to be a deciding factor. Nuclear power allowed the Coalition to run a campaign that accepted the common, and at the moment electorally unchallengeable, position that climate change is a problem, and presented voters with a solution that credibly differentiated the Coalition from the Labor Party.

The second claim about the election is that Donald Trump helped cause the Coalition’s loss. This has been championed not only by the Coalition’s political enemies, but also by many of its supporters, as well as journalists and analysts. This also doesn’t stand much scrutiny. Our respondents barely mentioned Trump. If he was discussed negatively, again it was from someone who was never going to vote for Dutton. Trump was also mentioned positively by some Nationalists who thought Dutton was a bit ‘Trumpish’, and they liked that he was. Potentially Dutton could have defined himself as the prime minister best equipped to deal with Trump, and initially some voters agreed with this assessment. However, as the campaign progressed and Dutton’s appeal slumped, Albanese came to be regarded as equally capable of managing the Australia/United States relationship.

The third claim is that the Coalition’s plan to prevent federal public servants working from home was a decisive element in the campaign. My respondents in fact barely mentioned it as an issue. This may be due to the nature of my sample. However, the relevance of the work from home issue is that Dutton’s decision to abandon the plan after a fierce public backlash was one of the reasons for his declining popularity. Dumping the plan raised questions about how well-thought-out his policies were and whether Dutton was strong enough to deliver them.

The original dichotomy between Dutton and Albanese was that Dutton was perceived as strong and Albanese was perceived as weak. In January we had Coalition supporters, who were the most likely to use the word ‘strong’, with comments such as ‘Mr Dutton does not prevaricate and offers a strong alternative which is credible’, or ‘A strong leader with experience in small business and policing, accepts good advice’.

By election weekend this ‘Nationalist’ voter had a very developed and not very complimentary view of Dutton:

The judgment was not just of Peter Dutton, but also of the Coalition, and from a voter whose preferences they needed.

On the other hand, the mood towards Anthony Albanese was probably best summed up by this comment from a Greens voter: