The National Party has dropped below 25 percent in the latest RNZ-Reid Research poll, while Opportunity has reached a new high and could decide the next government.
National has dismissed the result as "out of step" with other polls, just nine days before voting closes.
Christopher Luxon's party is on 24.8 percent, down 1.1 percentage points since the last RNZ poll three weeks ago. That puts National below its disastrous 2020 election showing, though still above its historic low in 2002.
Labour remains the most popular party on 30.6 percent, down 0.2 points, while Opportunity records the biggest gain, rising 1.7 points to 7.2 percent.

Click to view
The full poll results can be read HERE
14 comments:
It could be an outlier but why risk it? Party vote NZF or ACT or face a left wing government. TOPs policies greatly align with the Left. Their candidates have ties to the Left.
Winston has declared multiple times he won't go with Labour. Don't listen to those who tell you he will.
I totally agree with Janine about not party voting National. It's clearly a waste this time. But use the vote for Act, not Winston. Fool me once and all that. The memory of all those centre-right voters who trusted Winston in 2017 should still burn brightly.
The sample is essentially self-selected as a stratified sampling approach is applied to a population that has chosen to offer itself for that purpose. 3% error margin, my arse - that would be for a barrel of marbles from which the sampler extracts a truly random sample.
3 polls in 2 weeks with 3 inconsistent results. What is consistent is that the two major parties are bleeding votes to the minor parties.
National & Labour sleepwalking to becoming minor parties in 2032?
Only the deluded would believe anything our corrupt MSM reported.
Go woke like National into Labour Lite and the inevitable happens - down she goes! Sooner the better!
Sorry Barend - you’re wrong. Just because online panels are “self-selected” (as you say), doesn’t mean they’re not representative.
Ngaire Reid knows her stuff. She’s been providing field work services to market researchers for decades.
Her poll uses quota control and weighting to ensure the sample matches the census population demographics.
Sure, some people volunteer for panels so they can “have their say”. Good on them. But quota control and population weighting means their opinions don’t skew the overall sample.
Other polls such as Varian take a 50/50 approach, half from online panels and half from random-dialled mobile numbers.
You should stop looking for sampling errors and instead look at what the poll is saying, allowing for statistical significance of course.
Polls ask party vote preference - except for a small number which target individual electorates (hard to achieve these days).
The Reid poll shows a collapsing National party vote. And the fieldwork was at a time when the split-vote campaign started to gain traction.
It doesn’t necessarily mean disaster for the coalition. National won’t necessarily lose seats. Just list MPs.
The coalition won’t necessarily lose seats either. Just there will be more ACT and NZ First MPs and fewer National.
What IS highly significant is that Opportunity remains above the five per cent cut-off.
Which means Luxon will have to talk to them (did we believe him when he said he wouldn’t?).
It’s highly likely we won’t have a government on 8 November. Specials and overseas votes won’t be counted at that point, and they may make a difference.
Certainly, coalition negotiations will not have even started.
Save some popcorn for closer to Christmas.
The sample is drawn from a population who offered themselves for research purposes - a self-selected population.
Statistical significance as given in these polls is meaningful in the context of an exercise in which sampling is truly random, like marbles in a barrel. Not when you have self-selected samples, and subjects who may choose to not participate. or may lie. Many people use polls to wag the finger at the govt - "I'll vote for someone else if you keep doing what you are doing" - but return to their old voting habits when the chips are down.
Sure, there are plenty of statistricks (not a typo) that one can use to make it look like a kosher sample, but at the end of the day if you have a hand-drawn circle you don't calculate its area accurately simply by using pi to more decimal places.
Having said all that, polls do have some value, especially when looked at in sequence - they become more accurate predictors the closer they get to the 'big day'.
Polls are pretty useless but loved by the lazy, ignorant MSM as they’re a self generated story. They vary wildly - by more than their own stated margin of error. That alone tells you they can’t be relied on. The only useful thing about polls is any steady trend yet as this poll is wildly out of kilter with its own pattern one can only conclude it is junk. Either the methodology is flawed or they happened to get an unlucky sample that was in a mood to give the Nats a bloody nose.
PS to earlier reply:
>"Just because online panels are “self-selected” ... doesn’t mean they’re not representative."
I think they call that 'a'-about-'t' thinking in Aus!
If they're self-selected (or purposively selected by the researcher) we can no longer ASSUME that they are representative of the population from which they are drawn. When we take random samples, we ASSUME that they are because we have removed (hopefully) sources of bias,
Derek Cheng is worth reading in the Herald regarding the wide margins of Reid polls throughout the year. He does a useful comparison of this year's Reid polls and then wonders why the Herald chooses Reid instead of Roy Morgan whose polls also vary wildly.
All polls are self selected, people either agree to participate or they don’t. Just because you don’t like the numbers is not a valid reason so dismiss them. Facts don’t care about feelings.
Anon541, OF COURSE all polls involve self-selected samples to a degree. That's one of the reasons the ritualised 3% error margin is so much bunkum, as that is derived from barrels of marbles (figuratively speaking) which have no say in whether they participate and can't tell pollsters fibs.
I LOVE numbers, especially the tricky ones - am very experienced in sampling and analytical procedures and was often consulted by colleagues about them. What I dismiss is extrapolations from flawed samples and analyses to the populations they ostensibly represent. After all, the whole point of taking a sample is to infer attributes of the population from which they were taken.
Polls, polls, polls. Every day a new one, usually wildly out of kilter with the previous one.
Enough already!
Post a Comment
Thank you for joining the discussion. Breaking Views welcomes respectful contributions that enrich the debate. Please ensure your comments are not defamatory, derogatory or disruptive. We appreciate your cooperation.