Newshub are running a front page story with this headline:
Prime Minister's office apologizes for misleading comments about NZ's child poverty figures
It says "Jacinda Ardern has claimed poverty figures are improving but incorrectly suggested the data included children living in non-private dwellings".
PM Ardern's chief press secretary has since apologized, saying the data do not include this. Seems that Ardern, who is Minister of Child Poverty Reduction, is unsure what the child poverty statistics are measuring.
The Chief Press Secretary at the PM's Office has now stated, "The Prime Minister made an error at post-cab on Monday based on incorrect information I provided to her. The PM asked me to clarify if the Material Hardship Survey picked up people in non-private dwellings which I checked with officials who confirmed 'it doesn't'. However when relaying this to the Prime Minister... I accidentally relayed it as 'it does. I apologize for this error".
It goes onto say "There are around 4,000 children in emergency housing nationwide. Even if all these children were in material hardship in their current circumstances (and many wont be as their household housing costs are less than private rentals) the number is well within the margin of error".
Seems to me the PM's office may have issued a defense of the PM's misleading comments with a misleading statement. Why? Because the margin of error in the original survey was presumably based simply around the fact that the survey was based on a small sample size. In this case, all that one can infer from it is a statement like, "Of all of the children living in private dwellings, x per cent are in poverty, with a margin of error of y per cent".
That is, it is unlikely that the margin of error has much to do with the fact that the survey turns out to have been done from a non-random sample, which excluded whole chunks of the full population, like all of those in non-private dwellings.
To clarify, when a survey only samples - lets say several thousand families - then that small survey size introduces a "margin of error". However the issue Newshub have picked up on is different. Newshub have discovered there was a significant "sample selection bias" in the original survey that is pertinent to calculating reliable statistics on child poverty. The PM's stats are unreliable & the attempt of her office to claim that thousands of missing data points don't matter only serves to heighten mistrust.
Sources: