Last month I posted about the increase in the proportion of votes that were cast early (postal, pre-poll and remote) as opposed to those cast on election day. 55.7% of votes were cast early, with just 43.6% casting their votes on the day.
For today’s post I am looking at how those voters actually voted, and how it’s changed over time.
Starting with the two-party-preferred vote, applying my estimate of the Bradfield 2PP, the gap between election day and the early vote is the widest I’ve seen since at least 2004.
The Labor 2PP on election day was 5.3% higher than the early vote. The gap was 3.7% in 2022, and 5.1% in 2019. Such a large gap had the potential to overstate Labor’s position in the first hours of counting, but the counting of postal votes on election night, along with Labor’s convincing victory, seems to have made this not such a big issue in 2025.
Breaking down this vote by state,
The gap between the early vote and election day was at its widest in Queensland and then Victoria. Western Australia has the smallest gap amongst the states. The gap is reversed in the Northern Territory due to Labor doing very well in the remote vote.
When I break this data down by electorate, there are just five seats in Australia where the Labor election day 2PP was lower than on the early vote. Labor did 8.6% better on the early vote in Lingiari. Labor also did substantially better on the early vote in the large rural seats of O’Connor and Grey, presumably for the same reasons as Lingiari. The gap was much smaller in Durack and Mallee.
The seat with the biggest gap in the other direction was Macnamara, where Labor polled 56.8% of the early vote 2PP, and 68.0% of the election day 2PP. I suspect the Jewish vote in that electorate favours voting early, and has a conservative lean, in a way that isn’t relevant in the majority of seats. Macnamara was also the seat with the biggest gap in 2022, but much less than in 2025. Labor polled 58.7% of the early vote 2PP in 2022, and 66.6% of the election day 2PP. That’s a gap of 7.9% widening to 11.2%.
So most of the data that I’ve previously compiled on voting results by timing just covers the two-party-preferred vote, but for 2025 I have also looked at primary vote data for the main groups.
The Labor primary vote is almost the same on election day and in the early vote – but the Coalition primary vote is 6.4% lower on election day. The Greens vote and the others vote is substantially higher. So Labor’s big 2PP advantage on election day is not due to a higher Labor primary vote – it’s due to a higher vote for other groups that preferenced Labor above the Coalition.
Finally this map shows the 2PP result for the early vote and election day for each seat. While Labor won the 2PP overall in 100 out of 150 seats, they won 94 seats on the early vote and 109 seats on the election day vote.
Thanks Ben. What does this mean for the general narrative, based on polling trends, that Labor’s position got better and better as each week of the election campaign went on, including the final week? I see the gap was wider than in 2022, which might suggest the narrative is correct, but it’s basically the same as 2019 – and I don’t remember that election being one in which Labor’s position improved in the final two weeks – and you also show that this year, Labor’s primary vote was the same in both early and election day voting (is that itself normal or an anomaly?).
I have read about the Labor surge in the final two weeks as well as Labor winning hugely with undecided voters. This might have contributed to the wider early vote vs election day 2PP gap.
If you’re looking at 2CP numbers (if my maths is right),
1. The eventual winners lost the early day vote in Kooyong, Menzies, Deakin, Bendigo, Petrie, Bullwinkel and Forde (very narrowly).
2. In Melbourne, Wills and Bean, Labor won the early vote but not on election day.
Not sure if I have missed any other seats.
Older people vote early… Not surprising, if you’ve got time on your hands why not vote early?
Harder to do if you’re working full-time.
@Justin: I don’t think that it shows anything about the campaign. It’s about who choose to vote early & who choose to vote on the day; & who can only, by virtue of their conditions of employment, vote on the day or by virtue of their remote living circumstances cannot vote on the day?
One question to pose to the AEC (& through the AEC to the politicians) is given the ratio of votes why there are so few pre-poll booths & so many polling booths?
Another question is why don’t we close polling booths at 5.00pm. At my booth of 1150+ voters, in a working class area, not a single vote was cast after 5.00pm.
Some interesting stats I saw:
– All Tasmanian and South Australian seats had a majority of votes cast on election day. I am curious to see why early voting isn’t as prominent compared to other states.
– All Queensland seats had majority of votes cast early.
– Hinkler, Fisher which had early voting above 70%, all Gold Coast seats above 65%. All are likely due to the large retiree populations which @Julian has alluded to. Lingiari also did but it includes mobile voting which is pretty substantial.
– Some urbanised seats like Bowman, Fraser, Longman and Gorton had early votes around 65% and above. I’m pretty sure the Caroline Springs PPVC was the largest early voting place in Victoria.