Pasifika researchers are dismayed at the government's choice of a relatively cheap way of running the next census in 2023.
They and Māori say this is not the way to rebuild trust destroyed by the 2018 Census and they need to start looking for alternatives.
"We now are really sceptical and have little confidence in the value of the 2023 Census," Pacific Data Sovereignty Network coordinator Jacinta Fa'alili-Fidow said.
Stats NZ asked for between $226 million and $280m but has been given about $210m to $240m (the higher figures factoring in capital costs and contingencies) despite a business case that argued strongly against the "do-the-minimum" cheaper option.
The Pacific Data Sovereignty Network helped Stats NZ run consultations with Pasifika people, who made their preference clear for the department's first-ranked choice.
"The bare minimum does nothing for Pacific," Fa'alili-Fidow said.
It would not improve data that benefited everyone, such as on the overcrowding of households - which agencies confronting the pandemic had wanted information on but couldn't find when trying to plan which Pasifika might need to isolate away from their homes, she said.
Having been failed by Census 2018, people were increasingly sceptical of centralised research that appeared not to benefit them, Fa'alili-Fidow, who is also the chief executive of Moana Research, said.
"A lot of the concerns that we put forward around the digital-first approach have played out."
For 2023, work had begun, but more was needed in the in-between time.
"[But] we no longer have the confidence that all of the recommendations that we've made, will play out," she said.
"We do wonder, should we even bother?
"So it will mean that we'll probably be forced to find alternative data solutions and the census just will continue to have the reputation that it has - that is, that it is an unreliable source of good important data for the Pacific."
Stats NZ's business case states that the cheaper census options that were studied, would not only risk returning poor data quality but would not rebuild trust with Māori and Pasifika communities, undermining 2023 as a stepping stone to better future people counts.
Yet the government has chosen one of those options.
It has said Stats NZ had assured it that it could deliver a high-quality census for $210m; this option was recommended by Treasury, which rejigged its original advice to the government after Stats NZ suggested ways of using an extra $6m that boosted the number of Census collectors to almost 5000, well above 2018's and well below 2013's numbers.
Iwi risk becoming 'casualties' as census transforms
This alarmed Waikato University professor and leading Māori demographer Tahu Kukutai.
"There was a loss of confidence in the system and the ability of the system to deliver," she said.
"And that's a real shame because I've looked at census models all around the world ... and Stats NZ was often held up as kind of a gold standard in census taking and a gold standard in the production of indigenous statistics."
Stats NZ's consultation was sound, as was its intent good to fix the faults of 2018, but the outcome was a letdown, she said.
The chosen option's chance of meeting the future needs of customers for rich data was rated just two out of five in ministerial advice in February, released under the OIA.
"It's not just about 2023 - this is about the future of the census," Kukutai said.
"And the part that I think is particularly alarming for Māori, is that Stats NZ is engaged in an ongoing process of transformation and the risk is that iwi are going to be the casualties."
The risks around 2023 called into further question this decade-old transformation push to go digital and gather more census data online, especially since this went so badly awry in 2018.
"That's fine - but before you embark on that, you have to make sure that you're future-proofing that system so that it works for Māori," Kukutai said.
She was not advocating iwi breaking away from the census, but that they began looking at additional ways to gather the data.
"We want to be credible world leader in the trusted use of shared data - that's been a mantra that's been repeated time and again for some years now.
"And so, let's test that, let's break away from the traditional model and allow others, like iwi, to start developing and proposing and being resourced for models that actually work for them."
Similarly, Fa'alili-Fidow said Pasifika would keep working with Stats NZ, but if no change was forthcoming in a year people would be best to look at gathering data in their own way, such as through churches, she said.
Pasifika health researcher Dr Corina Grey said the government's approach did not generate trust and instead risked history repeating itself.
"The fact that these papers point out that there is a very real risk, particularly for our Māori and Pacific populations, and that they're prepared to take that risk, I think that is very disappointing."
It "absolutely" appeared they were cutting corners that could not afford to be cut.
"They've undertaken this consultation and then not taken up the advice of the people consulted," Grey said.
"We do need to constantly look at the decisions we make and make sure that it is populations that may be more marginalised, like Māori and Pacific, whose rights and needs are at the forefront of decisions."
She said there needed to be good governance around decisions like this.
"And really good advice - and there was good advice from the consultation, which the government has not chosen to take up."