Curia
32 polls · 3 Nov 2023 to 3 Jul 2026
- 1 of 32 without a sample size
Green−1.1
Others+0.7
Labour+0.5
TOP−0.5
National+0.4
ACT−0.1
NZ First+0.1
Te Pāti Māori+0.1
Polling
The pollsters
6 organisations have published party-vote polls for this election. They do not all poll the same way, they do not all report the same parties, and their figures do not all sit in the same place. This is what the trend is built from.
The bars below show how each house’s figures for a party have sat relative to the combined average at the same dates — to the right if above it, to the left if below. This is not a measure of accuracy: the average is not the truth, and a house that differs from it may be the one that is right. Differences like these are normal and reflect how each firm samples, weights and handles undecided voters. They are shown because they are real, and because a reader comparing two polls deserves to know the houses are not interchangeable.
Roy Morgan
32 polls · 12 Nov 2023 to 7 Jun 2026
- Reports every party, sample size on every poll
Labour−2.5
Green+1.7
National−1.0
ACT+0.9
TOP+0.8
Others−0.6
Te Pāti Māori+0.3
NZ First+0.1
Talbot Mills
19 polls · 5 Feb 2024 to 5 Jun 2026
- Never reports Others
- 8 of 19 without a sample size
- Includes an internal party poll
Labour+2.8
ACT−0.8
Green−0.6
TOP−0.5
National−0.4
NZ First+0.2
Te Pāti Māori−0.0
Verian
15 polls · 12 Feb 2024 to 15 Jun 2026
- Reports every party, sample size on every poll
National+1.7
ACT−0.8
Labour+0.7
NZ First−0.6
Te Pāti Māori−0.4
TOP−0.4
Green−0.3
Others+0.1
Freshwater Strategy
6 polls · 26 Nov 2024 to 8 Jun 2026
- Never reports TOP
Labour+2.5
Te Pāti Māori−0.9
National−0.8
Green−0.6
Others+0.6
ACT−0.5
NZ First+0.4
Reid Research
6 polls · 24 Mar 2025 to 5 Jul 2026
- Reports every party, sample size on every poll
Labour+1.3
Others−1.1
ACT−0.5
Te Pāti Māori+0.4
Green−0.4
National−0.1
NZ First+0.1
TOP−0.0
House effects are measured against the sample-weighted trend across all polls, which includes the house itself — so for the largest houses the real difference is a little wider than shown. Effects are computed only for parties a house has reported at least four times. A gap in coverage matters most near the 5% threshold: an average over a party that some houses never report rests on a different set of pollsters than an average over a party they all do.