Greens up on every measure in IPSOS polling

Additional data from the latest IPSOS Issues Monitor shows the Green Party gained public confidence on every concern the survey measures.

The pollster is warning against drawing too many conclusions - including whether it could signify an increase in party vote, or that the party is taking support from Labour.

The Greens say there are risks to political polling, but they're welcoming the results - and are calling for political parties to publicly release their internal polling in the interests of "equity".

The Issues Monitor survey questions New Zealanders every three months about what their three biggest concerns are - from cost of living, to health, crime, immigration and more - and which party they think is best able to tackle them.

The latest edition showed Labour remained most capable on seven of the top 10 issues, but with downward trends for the party on several of the top concerns.

On the top five, both major parties' ratings on their ability to manage the top concern - cost of living - was down; there was little change for either on managing healthcare; National was up and Labour down on managing the economy; both were down on fuel prices; and National was up and Labour down on housing.

The story was different for the Greens, with increases of 2 points, 2 points, 3 points, 1 point, and 4 points on the top five measures.

This was against the backdrop of declining support for the party in recent political polling - with a slow decline from January to May across a range of polls.

Since then things have changed somewhat, with Talbot Mills reporting a four-point bump for the Greens in June - doubling their support over two months. The latest Taxpayers Union poll showed a nearly 2-point increase; and a 1.5-point rise in Roy Morgan.

RNZ went back to IPSOS for the full numbers on all the parties' results, which showed the Greens gaining across every measure.

See the full results provided at the end of this article.

How robust are survey results?

IPSOS' New Zealand manager Carin Hercock told RNZ the statistically significant increases were the important ones.

"Statistically significant means that it's a change that is beyond the expected margin of variation within results," she said, "That there has been a genuine change in the sentiment."

The three- and four-point bumps on economy and housing were statistically significant, as were the party's gains on unemployment; poverty; immigration; race relations/racism; defence/foreign affairs/terrorism; crime/ law and order / violence and antisocial behaviour; taxation; and household debt.

The increases across the other measures were just one or two points.

She said in some areas it did look like numbers for confidence in Labour's ability to handle individual concerns were going down when the Greens' went up - but not in all of them.

"Immigration, the Labor Party dropped by one point, and the ... Greens went up by three ... it does appear on some issues there was a relationship in the last wave, but prior to the last wave Labour had actually been trending up on most several of the issues, so you know there's not a strict conclusion that you can draw."

Hercock said IPSOS was not a political poll and did not show levels of voter support for individual parties.

However, the survey used a very consistent sample design, and the fact there were variations in the changing support across the different issues indicated it was not a blanket group of people backing the Greens on every measure.

"There's no indication that it's the type of people who are responding to the survey, we consistently use the same sample quotas and have been doing so since the survey started.

"We run quotas based on age, gender, and region, and then we do some minor weighting across age, gender, region, and ethnicity."

She said census data showed how many people lived in different regions; their gender; age; and other population-level statistics - and where the sample of 1001 people surveyed did not quite match that, results would be adjusted slightly to represent those differences.

"But because we use weight quotas, we don't have to do a lot of weighting with the survey."

A sample size - the number of people polled - of around 1000 was common for big polling companies because that reliably produced results with a fairly small margin of error, which gave more confidence in the results.

She warned against drawing too many parallels between the Issues Monitor and political polls, and said although parties that had a significant lead on the top issues historically got a higher share of the vote, this did not necessarily translate to which would be able to form a government.

"It has to be quite a lead, I think is the key point there."

Some political polls were well established and had good historical data, and Hercock pointed to Research Association membership as a marker of reliability.

She pushed back on politicians who would dismiss polling - including political polling - as just a lot of numbers moving around.

"As researchers we're trying to give voice to the sentiment of the people, and if politicians choose to ignore that, well, so be it. But my view is that you're far better to have a deep understanding of what the population you represent is thinking at any given time," she said.

"There will be always be variation because there's always going to be variations in people's mindset - people don't actually often decide what they're going to do with the vote until the day that they're in that polling booth ... I think the general sentiment is definitely being collected by those pollsters."

Greens welcome survey results

Greens co-leader Marama Davidson told RNZ the IPSOS survey had "an effective method", and showed the party's policy message was "resonating".

"Our focus is very clear: people can have affordable lives, we can protect nature, we have enough to make sure that decisions are made for all of us, and we're going to keep putting those solutions out."

Overall, the quality of political polling was dependend on how it was carried out.

"I think it always comes back to what you're asking people, the methodology that you're using. I think we probably know by now that there's a generational difference between who uses landlines and who have never used a landline in their life."

She said polling had other limitations and difficulties to navigate - including that in some cases they could in some cases become self-fulfilling or self-perpetuating.

"Also I want to ask the question of equity across access to polling, you know, you've got some political parties who can afford to do it every day, you've got political parties like us who can afford to do it a few times a year."

One solution could be to have an independent publicly funded group that would carry out polling, she said - and she expressed an openness to having political parties' internal polling made public.

"Yeah, we'd certainly be up for that, absolutely," she said. "Again, this is just me sort of having thoughts, but there is a valid area for authentic and independent research to be used in a way that helps our democracy."