The poll of polls works by running a simulation fed with the most up-to-date polling data from multiple polls. We run the simulation for if the election were to be held this Saturday - and for if the election were to be held on October 14, when it is currently scheduled.
Running that simulation multiple times gives people a sense of how likely various outcomes might be as of right now. It does not, of course, factor in a future event that would upend the election (Jacinda Ardern taking over Labour in 2017, for example).
The Herald ran 8000 simulations of possible outcomes if the election were held this Saturday. Simulations were also run on the likely outcome of an election on the actual polling day, October 14. The simulation includes the results of two new polls, the Taxpayers’ Union-Curia of last week and the 1 News-Verian Poll, released on Monday night.
The poll of polls simulation gave National and Act, the not-so-affectionately named Coalition of Cuts, a 45.8 per cent probability of being in a position to form a Government on election night.
That’s just slightly ahead of the Labour-Green-Te Pāti Māori grouping (the equally not-so-affectionately named Coalition of Chaos), which has a 40.3 per cent probability of being able to form a Government.
Currently, Labour and the Greens have just a 2.1 per cent chance of getting over the line on their own. There’s a greater chance of a hung Parliament, which the poll of polls gives a 13.9 per cent probability.
If an election were held this weekend, National and Act would have a 41.7 per cent probability of winning it. They would have a 100 per cent chance if they were able to form a Government with Te Pāti Māori.
Labour, the Greens and Te Pāti Māori have just a 30.1 per cent chance of winning an election were it held this weekend, and Labour and the Greens alone have a zero per cent chance.
The chance of a hung Parliament were the election held this weekend is a relatively high 28.2 per cent.
How the ‘poll of polls’ works
The Herald’s “poll of polls” combines polling from different pollsters to predict the party vote for the 2023 election.
The model imagines that in any given week there exists an unobserved voting intention that is partially measured via opinion polls and is accurately measured once every three years by an election.
The model can also make predictions about how voting intention can evolve between now and election day.
This approach enables us to estimate each polling organisation’s accuracy in previous elections and then use that to inform our predictions.
The Herald has based its model on a New Zealand election forecast developed by statistician and data scientist Peter Ellis. Ellis developed the model in a private capacity before taking on his current role as director of Statistics for Development Division at the Pacific Community (SPC).
Ellis also used the model to forecast the Australian election. Similar models have been used to make predictions about the German and Swedish elections. Both Germany and Sweden have proportional representation electoral systems similar to New Zealand’s MMP system.
Our “poll of polls” combines results from a range of pollsters, who are signatories to the NZ Political Polling Code, specifically Curia, Kantar Public, Talbot Mills and Reid Research.
Details of the model, including the source code, are available here. Between now and the election we will be looking to improve some aspects of the model, in particular the handling of polls which do not provide polling data for some of the smaller parties.
Currently, in order to be included in the model, a party must have polled over 2.5 per cent in at least three polls.
It is assumed that parties currently holding an electorate seat retain them and no other parties win an electorate seat. Future versions will enable readers to modify this assumption.
Thomas Coughlan is deputy political editor of the New Zealand Herald, which he joined in 2021. He previously worked for Stuff and Newsroom in their Press Gallery offices in Wellington. He started in the Press Gallery in 2018.