Are learners attending, learning and leaving with a next step?
What the media told New Zealand about education: the issues, the fixes on the table, and how each party’s policies measure up
We analysed 3,606 education headlines published by 124 outlets between June 2023 and September 2026. The coverage made reform, and resistance to it, highly visible: a rewritten curriculum, NCEA’s replacement, attendance crackdowns, cheaper school lunches, charter schools, teacher strikes and the break-up of Te Pūkenga. What it rarely showed was whether any of it helped students learn. Only about one education headline in ten used any outcome language at all.
3,606
education headlines, 4.2% of the record, making it the sixth most reported topic
9.5%
of education headlines used outcome language, falling to 2.7% for charter schools
5 of 252
school-lunch headlines, at most, connected lunches to learning, attendance or achievement
33.7%
of education headlines were paywalled in 2026, up from 21.9% in 2023
The issues and the solutions
Each section sets out the problem, what has been done, what the coverage left unanswered, and what would show whether a fix is working. Click a heading to open it.
1. Curriculum, literacy and numeracy
The problem
Falling achievement in reading, writing and maths drove a rewrite of the curriculum, a shift to structured literacy and phonics checks, and a maths action plan. By late 2025 the story had become one of implementation strain: hundreds of principals urged the Minister to halt a “rushed” rollout (RNZ, October 2025), and in 2026 dozens of principal and teacher associations signed a letter opposing the changes.
What has been done
The curriculum was rewritten and phonics checks were introduced. Early results arrived in May 2026: the Herald reported encouraging maths results, while RNZ reported that fewer than half of five-year-olds passed the phonics check after a year at school. The Post reported errors in AI-assisted curriculum drafting in July 2026.
What we still don’t know
Whether the new curriculum is improving learning, and for whom. Early results were never connected to a baseline, a target or a follow-up. The one headline that tested a success claim came at the very end of the period: The Post reported that the Government’s claim of “transformational” literacy gains rested on data from 14% of schools.
Solutions, and how to tell if they work
On the table: continue the rollout, or pause it and create an independent body to oversee curriculum and assessment. The test: track announcement → teacher preparation → classroom use → assessed learning, with results reported for the students previously left behind.
2. NCEA: from declining results to replacement
The problem
Falling results, failed online exams and a struggling literacy and numeracy co-requisite. RNZ reported in December 2024 that 45% of students had failed that year’s online literacy and numeracy tests, and ERO found the revised Level 1 was not working. That NCEA had a problem was widely accepted.
What has been done
The Government announced NCEA’s replacement on 5 August 2025, the sharpest single-issue burst in the education record (54 headlines between 5 and 15 August), and confirmed it in March and May 2026. Nearly 90 principals pleaded for it to stop a “politically driven” replacement.
What we still don’t know
How the replacement will fix the problem, and which students are at risk. The contest over outcomes began before the new system arrived: a survey found most teachers wanted NCEA improved, not removed, and a PPTA report projected that the replacement would fail more students, which the Herald put at 14,930 at Level 2. The Minister called that a “defence of mediocrity”. Good free analysis existed, but readers had to assemble it across outlets.
Solutions, and how to tell if they work
On the table: proceed with the replacement, or pause it. The test: qualification results and the number of unqualified school leavers, reported for the same student groups before and after the change.
3. Attendance: presence or participation?
The problem
In October 2024 ERO found chronic absence had reached “crisis levels”. Curriculum reform cannot help a child who is not at school.
What has been done
Attendance reporting and enforcement were tightened, including prosecution of parents. Term-two attendance reached a three-year high in 2025, and Newsroom reported record attendance in August 2026.
What we still don’t know
How much of the improvement came from enforcement, from less illness or from better engagement. One in four schools was struggling to meet the Government’s goal, there was little evidence the truancy fund had worked, and record attendance masked a persistent gap for Māori students. Attendance was the best-measured school topic (26.2% outcome language), but causes such as illness, poverty, transport and disability featured in only 15 headlines.
Solutions, and how to tell if they work
The test: target → school response → barriers addressed → regular participation → attainment. Report regular attendance alongside its causes, not just its rate.
4. School lunches: contracts or learning?
The problem
The free school lunch programme exists so hungry children can learn. It was redesigned as a cheaper model in May 2024 and contracted to a consortium led by a global catering company.
What has been done
The February 2025 rollout produced 47 lunch headlines in a month: late and missing deliveries, burnt meals, “halal-friendly” meals containing ham, and an MPI investigation. Compass was later excluded from the next round. In July 2026 the Auditor-General’s review found only half the meals served in 2025 met health and nutrition standards.
What we still don’t know
Whether lunches improve learning, attendance or wellbeing. Coverage scrutinised procurement, cost and food safety closely: 35 headlines on contracts, 29 on cost and 23 on food safety. At most five connected lunches to the purpose the programme was meant to serve.
Solutions, and how to tell if they work
The test: procurement → delivery quality → uptake → nutrition and safety → learning and wellbeing. Judge the programme on whether children eat and learn, not only on price.
5. Charter schools
The problem
Charter schools were re-established with $153 million of Budget funding to offer choice and better results, despite Cabinet papers warning of higher costs and marginal benefits (RNZ, May 2024).
What has been done
The legislation passed in September 2024 and the first schools were announced. Converting state schools proved tougher than expected: the Kelston Boys’ High School bid was abandoned. In September 2026 two former charter schools were under scrutiny for growing debt and illegal board appointments.
What we still don’t know
Whether charter students learn more than comparable students elsewhere. RNZ reported in July 2026 that no independent testing was planned. Only 2.7% of charter-school headlines used outcome language, the lowest of any topic.
Solutions, and how to tell if they work
On the table: keep and expand the model, or abolish it again. The test: independent, comparable results for charter students against similar students in state schools, and published financial accounts.
6. Māori education and the Treaty
The problem
Achievement gaps for Māori and Pacific learners persist. Coverage centred on authority and content rather than learning results: Māori and Pacific learners were removed from extra funding priorities (June 2025), Māori words were cut from future junior readers (August 2025), and Treaty obligations were removed from school boards despite warnings to ministers.
What has been done
In December 2025 the Minister said the shift from a Treaty focus to closing the achievement gap was “paying off”. In May 2026 Labour pledged to restore the commitments.
What we still don’t know
If closing the achievement gap is now the measure, it should be reported against, and the headlines contained few such measures. One exception, that kaupapa Māori students were more likely to gain NCEA merit and excellence endorsements (RNZ, October 2024), received little follow-up.
Solutions, and how to tell if they work
On the table: an achievement-gap focus, or restoring Treaty obligations. The test: attendance, literacy, numeracy and qualification results for Māori and Pacific learners, reported regularly against a baseline.
7. Teachers and the workforce
The problem
Every reform depends on the people who deliver it. 2025 was the peak of industrial action: Ministry staff struck for the first time in two decades, 2,300 learning-support staff walked out, and secondary teachers struck over an “insulting” 1% pay offer. Northland schools rostered students home, and in 2026 the PPTA warned that booming rolls were fuelling a teacher shortage.
What we still don’t know
Whether the workforce has the capacity to deliver curriculum, NCEA, attendance and learning-support changes all at once. Pay disputes and reform programmes were reported side by side, but seldom as one question of capacity.
Solutions, and how to tell if they work
Any reform should say which teachers, aides and specialists will deliver it, where they will come from and what else they are being asked to do. The test: report workforce supply alongside each reform that depends on it.
8. Behaviour, safety and learning support
The problem
Stand-down rates were the highest in 20 years in 2024, and student assaults hit a record high in 2026. Learning support drew only 55 headlines, but they were the most alarmed of any topic (20.0%): more than 5,000 children were waiting for specialist support in October 2025, and demand blew past the Government’s forecast in September 2026.
What has been done
A cellphone ban, which a 2025 review found was working; a $646 million learning-support boost (May 2025); and ACT’s 2026 proposal to let teachers use “reasonable force”. Teachers responded that rising assaults needed support, not force.
What we still don’t know
Why assaults and stand-downs are rising, and whether the new learning-support money has reduced waiting times. Funding announcements and shortfalls appeared as separate stories.
Solutions, and how to tell if they work
The test: identified need → assessment → funded support → access → progress. Report learning-support waiting times and behaviour data every year.
9. Early childhood education
The problem
Fees, pay parity, ratios and a regulation review. Early childhood coverage was 39.4% paywalled, the second-highest share of any education topic.
What we still don’t know
Two headlines sum up the gap. Just 153 families were receiving the maximum FamilyBoost credit (Herald, July 2025), and officials admitted there was no evidence behind childcare staffing ratios (Newsroom, July 2026). One is a policy measured only by uptake; the other is a safety standard never measured at all.
Solutions, and how to tell if they work
The test: affordability and safety outcomes for families, not just the number receiving a rebate.
10. Universities, polytechnics and the step into work
The problem
Universities cut jobs and courses under financial strain; two were rated high risk in 2024. Te Pūkenga was dismantled and ten polytechnics returned to regional governance. By the end of the period, apprenticeship numbers had hit a 15-year low and graduates were being “pushed, not pulled” to Australia.
What has been done
Regional polytechnics were restored, fees-free was scrapped, and a $230 million Waikato medical school was committed. In the 2026 campaign, National proposed lower loan repayments for graduates who stay in New Zealand, and Labour proposed writing off 10% of some balances.
What we still don’t know
Whether regional access, programmes and completion survived restructuring. Te Pūkenga coverage stopped once the decision was made: no headlines in 2026. Vocational education was the most paywalled topic (47.7%) and among the least measured (3.4%). The apprenticeship and enrolment figures were never connected to the restructuring.
Solutions, and how to tell if they work
The test: funded place → enrolment → completion → destination → employment and earnings, by region and field of study.
How the parties’ policies measure up
We used the same six-point test for every party. A good education policy should:
1. Define the problem, the learners affected and the stage. 2. Show the pathway from policy to learning. 3. Resource the workforce. 4. Measure participation, not only presence. 5. Protect equity and continuity through structural change. 6. Make delivery and outcomes accountable.
National, ACT and Labour are assessed on their policies as reported in headlines up to 29 September 2026. New Zealand First (7 headlines), the Greens (5), Te Pāti Māori (2) and The Opportunity Party were barely visible in that coverage, so their policies come from party announcements, policy platforms and news reports published up to early November 2026; sources are listed in each section.
National
Policies as reported: a rewritten curriculum with structured literacy and a maths plan; NCEA’s replacement; a school cellphone ban; removal of Treaty obligations from school boards; a new agency for school property; the break-up of Te Pūkenga; an end to fees-free; and, in the 2026 campaign, lower loan repayments for graduates who stay and a ban on screens in early childhood education. Erica Stanford led the curriculum, NCEA and Treaty changes.
Pros
It acted on problems that were widely accepted: falling literacy and numeracy, and a failing NCEA. Early measures exist, with encouraging maths results reported in May 2026. The cellphone ban is one of the few reforms with visible follow-through: officials said the evidence was “marginal” in 2023, and a 2025 review found it was working. The loan-repayment policy is aimed directly at the graduate brain drain.
Cons
Hundreds of principals and dozens of sector associations opposed the pace of the curriculum rollout, and teachers were striking at the same time, which leaves workforce capacity (test 3) unaddressed. Its “transformational” literacy claim rested on data from 14% of schools. Fewer than half of five-year-olds passed the phonics check. Treaty obligations were removed despite warnings, and Māori and Pacific learners were dropped from extra funding priorities, which raises the equity question in test 5. Te Pūkenga’s break-up was never followed by evidence on regional access, and apprenticeships hit a 15-year low.
Against the test: clear on problem (1); weak on workforce (3) and equity (5); success claims thinly measured (6).
ACT
Policies as reported: a cheaper school lunch model; re-established charter schools; tougher attendance enforcement, including prosecution of parents; and, in 2026, a proposal to let teachers use “reasonable force” with disruptive students. David Seymour led all four.
Pros
It treats being at school as the first step to learning, and attendance has risen to a three-year high and then record levels. Attendance was the best-measured school topic. The reasonable-force proposal responds to a real and growing problem: record student assaults. Charter schools offer families a choice.
Cons
The lunch rollout failed publicly, and the Auditor-General found only half of 2025 meals met nutrition standards. At most five of 252 lunch headlines connected lunches to learning. Cabinet papers warned that charter schools would cost more for marginal benefit, no independent testing is planned, and two former charter schools face debt and governance problems. It is unclear whether attendance gains reflect engagement or enforcement, and there was little evidence the truancy fund worked. Teachers said assaults need support, not force.
Against the test: clear on problem (1); counts presence more than participation (4); weak on pathway to learning (2) and on accountability for charter-school results (6).
Labour
Policies as reported: pause the curriculum overhaul and NCEA changes; create a body to oversee curriculum and assessment; reinstate climate change in the curriculum; restore Treaty obligations for school boards; abolish charter schools again; and write off 10% of some student loan balances.
Pros
The pause responds directly to the implementation concerns raised by principals and teachers, and a survey found most teachers wanted NCEA improved, not removed. An oversight body speaks to the accountability gap in test 6. Restoring Treaty obligations addresses the equity concern in test 5. The loan write-off, like National’s policy, targets the graduate brain drain.
Cons
What a pause would replace the reforms with was never visible in the coverage, nor were alternative plans for achievement, attendance or the workforce. Labour was largely absent on education in 2025 (nine headlines); in the year NCEA’s replacement was announced, its most concentrated coverage was its spokesperson not responding to briefing invitations. Abolishing charter schools would be another structural change for the students in them, which test 5 asks to be managed. The cost of the loan write-off was not reported.
Against the test: responsive on equity (5) and accountability (6); unclear on pathway to learning (2) and cost; visible mainly as reaction until the campaign.
New Zealand First
Policies as announced: NZ First has released comparatively little new education policy for the 2026 campaign. Its platform proposes linking outcome goals across early childhood, school, tertiary and employment; an Independent Curriculum Council to advise on curriculum development; and a focus on mastering reading, writing, arithmetic and science before moving on, with priority for children with learning difficulties, including Reading Recovery. For early childhood: development targets for motor, social, language and numeracy skills, ratios based on group size rather than whole-centre averages, unannounced spot checks, and keeping 20 free hours. For training and tertiary: support for cadetships, apprenticeships and second-chance providers; regional polytechnics; matching loan repayments for graduates who stay; and a higher threshold before loan repayments start. Parents would see relationships and sexuality education content, with an opt-in or opt-out. Separately, a $5,000-a-year Kiwi Kids Grant for each of a family’s first three children, for their first three years, framed as a birth-rate incentive. In government, Winston Peters announced the end of fees-free in May 2026, saying the money should go to trades.
Pros
Linking outcome goals from early childhood through to employment is the clearest statement by any party of the learner pathway the report found missing (tests 2 and 6). An independent curriculum council answers principals’ complaints about rushed, politically driven change. Group-size ratios and spot checks respond to the finding that childcare ratios rested on no evidence. Priority for children with learning difficulties targets the most alarmed, least covered area in the report.
Cons
None of these policies has published costs (test 4), including the Kiwi Kids Grant, which works out to $15,000 per child. The Office of ECE noted limited engagement with the sector’s current conditions, and CCS Disability Action found no education policies to assess. Much of the platform is a list of intentions without targets or timelines. As a governing party, NZ First shared responsibility for the reforms whose pace it now criticises, and ended fees-free without a reported plan for the students affected.
Against the test: strong on pathway design (2) and early childhood accountability (6); uncosted (4); low campaign visibility.
Sources: AEC NZ policy summary, 9 Sep 2026; Office of ECE guide, Nov 2026; RNZ via Scoop, 9 May 2026; CCS Disability Action, Oct 2026.
Greens
Policies as announced: restore the original Ka Ora, Ka Ako lunch programme with local providers and extend it to 400 more schools and 150,000 more students, at about $6.50 a meal, costed at $472 million in the first year and rising to $602 million by 2030/31. Free early childhood education from 2029, with fees capped at $10 a day before then and public funding ended for for-profit centres by 2028. A $380 weekly payment, plus a $1,500 relocation grant, for student teachers on compulsory placements, costed at about $75–80 million a year. Wider criteria for learning support, with a single agency to coordinate it. A long-term workforce plan co-designed with the sector. Opposition to charter schools and to scrapping NCEA, and a commitment to fight for the return of fees-free.
Pros
The lunch policy deals with quality as well as reach, after the Auditor-General found the current scheme’s performance was not being properly tracked. The biggest policies come with costings (test 4). Paid placements and a workforce plan speak directly to teacher supply (test 3). A single learning-support agency responds to the waiting lists and fragmented support identified in the report.
Cons
No specific funding source is named beyond “a fair tax system”. David Seymour says the current lunch scheme gets “the same results as the old programme, but cheaper”; neither side has shown what lunches do for learning, the gap the report found. Ending public funding for for-profit ECE centres is a large structural change, which test 5 asks to be managed without disrupting families. The sources we found did not set out an alternative plan for NCEA’s acknowledged problems.
Against the test: costed on lunches and placements (4); strong on workforce (3); light on the pathway to learning (2) and on measures of success (6).
Sources: 1News, 20 Aug 2026; Green Party release, 16 Oct 2026; Office of ECE guide, Nov 2026; CCS Disability Action, Oct 2026.
Te Pāti Māori
Policies as announced: a Māori Education Authority to oversee a Māori-led education system and direct more investment to kōhanga reo, kura kaupapa Māori, wharekura and wānanga. Te reo Māori as a core subject to Year 10, and mātauranga Māori embedded in the curriculum. A Kōhanga Reo Act with secure baseline funding and capital for new and refurbished kōhanga. School lunches extended to all kura and public schools. Free counselling in schools, and free public transport for students. A Māori-led taskforce on how disabled and neurodiverse Māori students are taught and supported. A universal student allowance at double the current rates; the living-cost portion of all student loans written off; and full loan write-offs for graduates who work in New Zealand for five years.
Pros
It names the learners and the problem (test 1) more directly than any other party: the persistent gap for Māori students that the report found was rarely measured. The disability taskforce targets the learning-support area the report found most alarmed and least covered. The five-year work condition on loan write-offs is aimed squarely at the brain drain.
Cons
The party has not published costings; it says these policies would be funded through its proposed tax system. The Taxpayers’ Union, a lobby group, estimated the living-cost write-off at about $4 billion and the universal allowance at $770–790 million a year. How the Māori Education Authority would sit alongside the Ministry, and what results it would report, is not yet set out (test 6). Published figures for kōhanga funding differ between sources.
Against the test: strong on problem and learners (1) and equity (5); uncosted (4); accountability unclear (6).
Sources: Waatea News, 3 Aug 2026; CCS Disability Action, Oct 2026; Office of ECE guide, Nov 2026; Massive Magazine.
The Opportunity Party
Policies as announced: a 10-year cross-party education plan, drafted with teachers and communities, that sets goals, assigns responsibility for delivery and tracks progress. Early childhood quality standards: minimum ratios (1:3 for under-twos, 1:5 for two-year-olds, 1:8 for three and over), Australian space standards, maximum group sizes, 80% qualified staff, and pay parity with school teachers. Civics taught from Year 6. Better pay and accredited professional development for teachers. Specialist support, such as speech-language therapists, psychologists and learning-support coordinators, available to every school. Less classroom screen time. A Citizen’s Income of up to $370 a week that would replace the student allowance.
Pros
Its central idea, a long-term cross-party plan with goals, owners and tracking, goes straight at the report’s main finding: reforms reported one stage at a time, with no one visibly accountable for the result (test 6). Its ratio and space standards respond to the finding that childcare ratios rested on no evidence. Specialist support in every school addresses the learning-support waits.
Cons
Costs are not set out on its education policy page, and Newsroom reported ministers dismissing the policy as an “unaffordable dream”. Many school measures are broad statements of intent (“pay qualified teachers well”). We found no position on NCEA, the curriculum rewrite or attendance, the biggest school issues of the period.
Against the test: strongest of any party on accountability design (6); good on workforce (3); light on specifics and cost (4).
Sources: Opportunity, Future Fit Education; Opportunity, 10-year plans; Office of ECE guide, Nov 2026; CCS Disability Action, Oct 2026; Newsroom, 3 Sep 2026.
New Zealand does not lack education reforms or education news. What it still needs is continuity: between reform and teaching capacity, between participation and learning, and between leaving school and a viable next step.
About this analysis
Source: the Social Herds headline record of 84,950 headlines from 1 June 2023 to 29 September 2026, built from the Democracy Project’s daily news briefings. The education subset is the 3,606 headlines with Education as the parent topic. Topics within it were identified by keyword search and overlap, so counts should not be added together. The unit of analysis is the headline, not the full article; counts measure visibility, not readership or the quality of reporting. “Outcome language” is a keyword search for results, pass rates, achievement, data, evaluation and similar terms. Party and minister visibility is based on name-matching and should be treated as directional. This report does not incorporate official achievement or attendance statistics; figures quoted are as reported in headlines.
The full report, including charts and the strand-by-strand table, is available as a PDF download.