Strip away the election modelling and take one latest published result from each polling organisation: Opportunity remains below Parliament’s 5% threshold. Excluding Roy Morgan, its average falls to 4.3%.
The New Zealand Herald’s latest “Poll of Polls” says the Opportunity Party is more likely than not to enter Parliament.
The Herald–Motu model gives Opportunity a 56.1% chance of making it and an average predicted election-night vote of 5.3%.
That may prove to be a good forecast. But it is not a simple average of the published polls, and those two things should not be confused.
Most readers hearing the phrase “poll of polls” would reasonably expect to see what the polls themselves average. So B2B News ran that calculation.
B2B News · Raw polling average
Opportunity polling, without the crystal ball
One latest result per polling organisation, equally weighted. The Herald–Motu forecast is shown separately so an average of polls is not confused with an election prediction.
B2B’s unadjusted average is 4.8 percent across the latest result from 5 polling organisations. The Herald–Motu model separately predicts 5.3 percent and a 56.1 percent chance of entering Parliament. Roy Morgan is the only pollster’s latest result at or above the five percent threshold.
What the Herald published
These are outputs from the Herald–Motu forecasting model, not individual poll results and not inputs to the B2B calculation. Read the Herald report.
View the underlying poll sources (6)
| Polling organisation | Fieldwork | Sample | Opportunity | Published source |
|---|---|---|---|---|
| Talbot Mills | 10–19 Jul | Not public | 4.7% | Spinoff report · Herald report |
| RNZ–Reid Research | 2–9 Jul | 1,000 | 4.7% | RNZ report |
| Taxpayers’ Union–Curia | 1–5 Jul | 1,000 | 3.3% | Commissioner release |
| Roy Morgan Excluded | 25 May–21 Jun | 891 | 6.5% | Pollster release |
| 1News–Verian | 13–17 Jun | 1,001 | 4.6% | Commissioner report |
| The Post–Freshwater Strategy | 5–11 Jun | 1,038 | Not reported | Pollster record |
Source notes: Private client poll; complete original tables and sample disclosure were not publicly accessible. Opportunity was not separately reported, so no value is inferred from the “other” total.
We took the latest available result from each polling organisation whose fieldwork ended within the previous 60 days. Each organisation received one equal place in the average. We applied no historical corrections, house effects, trend estimates or predictions about what voters might do on November 7.
The five available Opportunity results were Talbot Mills at 4.7%, RNZ–Reid Research at 4.7%, Taxpayers’ Union–Curia at 3.3%, Roy Morgan at 6.5%, and 1News–Verian at 4.6%.
Their arithmetic average is 4.76%, which rounds to 4.8%.
That is below the 5% party-vote threshold required to enter Parliament without winning an electorate seat. The Electoral Commission confirms that a party must receive at least 5% or win an electorate.
Why Roy Morgan changes the picture
Roy Morgan is the only organisation’s latest result to put Opportunity over the threshold, at 6.5%.
It also recorded Opportunity at 6% in its preceding survey. That means the two polls above 5% mentioned in the Herald article are not results from two different polling companies. They are two consecutive releases from Roy Morgan.
Remove Roy Morgan from the B2B News average and Opportunity falls to 4.3%.
We are not saying Roy Morgan should automatically be discarded. Its latest poll surveyed 891 electors by landline and mobile phone over almost four weeks, and its published result is a legitimate data point.
That is why our default average includes it. But readers should be able to see how strongly that single polling organisation affects the result.
What do the latest published polls average before anybody tries to predict the future? Today, the answer is 4.8%.
A forecast is not an average
The Herald is entitled to construct an election forecast.
Its published explanation says the model combines public and private polling, historical performance going back to 2014 and 4000 simulated elections. That is substantially more complicated than adding up current poll results and dividing by the number of pollsters. The Herald has itself described it as a predictive model.
There is also an interesting arithmetic detail.
The five polling results singled out in the Herald’s latest story—Roy Morgan at 6% and 6.5%, Verian at 4.6%, Reid Research at 4.7% and Talbot Mills at 4.7%—average to exactly 5.3%.
That does not prove the Herald’s model is a simple average. The Herald says it is not. But it demonstrates how counting two results from one polling organisation, rather than taking one latest result from each, can materially change the headline.
Our approach is deliberately less ambitious. It answers one question only: what do the latest published polls average before anybody tries to predict the future?
Today, the answer is 4.8%.
The necessary health warning
A 4.8% average does not prove Opportunity will miss Parliament, just as a 5.3% forecast cannot guarantee it will enter.
Polls are snapshots, carry sampling and other uncertainties, and can move considerably during an election campaign. The difference between 4.8% and 5% is too small to support a confident election-night conclusion.
The Research Association New Zealand’s political polling guidance stresses transparent reporting of samples, methods and uncertainty.
That is why B2B News will publish every input, every source, the calculation used and the time the data was last checked.
No crystal ball. Just the polls.
Join the discussion
Add useful context, ask a good question, or challenge an idea — keep it specific and respectful.