Are public services really getting worse?

And why the media won't let numbers speak for themselves.

According to The New Zealand Herald, an opinion poll commissioned by the Public Service Association (PSA), the trade union that represents public servants, “more than two-thirds of New Zealanders think the quality of the public service is either no different or has worsened under the coalition”.

The Herald’s hard copy put the boot in: “Faith in public service drops”, it declared. “Union-backed poll finds Kiwis think sector worse under coalition despite strong election odds”.

That doesn’t reflect well on the services that PSA members provide, but the targets were the current elected leaders (Luxon & Co) whom they oppose. Normally, moreover, the Herald gets accused of bias towards the centre-right, but on this occasion their interpretation of the survey weighed against the coalition government.

Using the same survey results, the headline could have read: “Most Kiwis don’t believe public services are getting worse”.

That is, a minority, only 38%, said that the overall quality of public services had become somewhat worse or much worse under the National-led government. A majority (54%) said public services were no different or even better.

Both headlines frame the survey in a biased manner. Here are the numbers themselves, then:

The balance of opinion was more towards “worse” (38%) than “better” (25%), but the modal response was “no different” (29%). The slice on which you spread the butter that tastes “neither better nor worse” makes a big difference.

The problem here is biased interpretation of survey results in news headlines. Reporting a survey in a manner that would allow the numbers to speak for themselves, and readers to make up their own minds, wouldn’t grab attention. “Faith in public service drops” has impact.

We also need to think pragmatically about how surveys work.

If I’d been asked by a surveyor whether public services were better or worse under the present government, then, depending on my mood at the time, I may have said “somewhat worse”. But on what grounds would I have made that judgement? Even as a person who follows public affairs, my evidence – weighed up in a second or two – is flimsy. It’s influenced by what I’ve read in the media (for example about public servants being laid off) and a generally negative vibe that things are going downhill and the budget is tight. Responses are also affected by political affiliation. Left-wingers feel more inclined to say “worse”, just because.

This survey wasn’t a genuine effort to understand how New Zealanders evaluate their public services. It was just fishing for anti-coalition sentiment.

The Herald used the verb “drops”, even though there’s no series of surveys – and hence no before-and-after comparison. For a long series, in contrast, there’s the Kiwis Count survey, curated by the Public Service Commission, an agency of central government, not to be confused with the PSA. Its results indicate that overall trust in the public sector has risen from 58% in December 2023, not long after the change of government, to 64% in March 2026. Back when Chris Hipkins was PM in 2023, the results varied from 58 to 61%.

Based on that survey, today’s headline might be: “Faith in public service rises”.

So which is it? Dropping or rising?

Readers are advised to look at all opinion surveys through a critical lens. But an important feature of the Kiwis Count survey is that it includes a question on service satisfaction. In the March 2026 quarter “84% of respondents were satisfied or very satisfied with their most recent government service, unchanged from the last quarter”.

Satisfaction with actual service delivery is much higher than general trust in the public service. That’s a standard finding in such surveys: we express greater confidence in those public services of which we’ve had close or recent experience.

Just asking a random person to make an instant judgement on whether things have largely gotten better or worse is – if you think about it – not a good way to gather valid evidence about real-world conditions. All you get is mood music. It’s better to ask people about what they know from their experience, for empirically grounded results.

By the way, the Kiwis Count survey also reports that “people who used exclusively digital channels for their most recent service had higher levels of satisfaction (88%) than those who used non-digital (82%) or mixed channel services (82%).” In my experience too, e-government is working quite well.