What does NZQA require for internal moderation of NCEA assessments?

What counts as internal assessment under NCEA?

NCEA runs across Years 11 to 13 at Levels 1, 2 and 3. Every achievement standard is either internally assessed by the school or externally assessed by NZQA. Internal standards are marked by the teacher who taught the course, against the standard’s Achieved, Merit and Excellence criteria, with Not Achieved where the evidence does not meet the standard. External standards go to NZQA through examinations and portfolios, and no amount of school moderation touches those grades.

The split matters because an internal grade is the school’s own claim. A school assesses against national standards under a consent to assess, and each internal result it reports says a student has met a national standard. NZQA does not re-mark every internal grade. It relies on the school’s quality assurance, checked by sampling afterwards. Students draw on classroom teaching, their own reading and study guides, and the school has to be able to show the graded work is genuinely the student’s own.

Junior classes in Years 9 and 10 sit outside NCEA, though many departments use the same criteria language early so students reach Year 11 already fluent in it. That is good practice rather than a requirement, and it carries no moderation obligation.

What does NZQA require for internal moderation?

NZQA sets out quality assurance that a school must undertake each year, for each standard it assesses. It has two halves: check the task before you use it, and check the grades before you report them. Neither half is optional, and neither is satisfied by a head of department signing a form at the end of the year.

The requirements are specific enough to audit against.

• All assessment material is critiqued and re-contextualised or modified as necessary, before use.
• For each standard, a sufficient sample of each teacher's grade judgements is verified.
• Verification happens before results are reported to NZQA.
• Assessment materials and a selection of student work are retained for external moderation.
• All advice from external moderation is actioned before the materials are used again.
• Evidence of internal moderation, and the school's monitoring of it, is retained until the end of the following academic year.

How do you choose which student work to verify?

NZQA is explicit that work should be strategically selected rather than selected randomly, including samples around grade boundaries and any grades that need review. In practice that means the A/M and M/E boundaries, where professional judgement varies most, plus anything a teacher has flagged as difficult. A random pull of six scripts will mostly return clear-cut cases and tell you very little.

Two details in the wording get missed. The requirement is a sample of each teacher’s grade judgements, so a department running one standard across four teachers needs a sample from all four, not one pooled set. And verification must be undertaken by a subject specialist with standard-specific knowledge. A deputy principal without the subject cannot do it. Small schools with a single teacher in a subject usually arrange verification with another school or through a subject association.

NZQA describes the requirement as a sufficient sample without prescribing a fixed number, so the defensible position is one you can explain: how many pieces, chosen on what basis, and what changed as a result. Record the grades that moved after verification. A moderation record showing that nothing ever changes invites the question of whether verification is real.

What does external moderation actually ask for?

External moderation gives assurance that assessment decisions are consistent nationally. Schools submit the task and supporting materials, the assessment schedule, and selected student work through NZQA’s external moderation application. Submission is due within 10 days of submitting results to NZQA, and there is an annual deadline in late October so schools get feedback before the next assessment year.

Sample requirements are set. For standards where Not Achieved, Achieved, Merit and Excellence are available, six samples are submitted: one at each grade, plus two more drawn from Achieved, Merit or Excellence, with a maximum of two at Not Achieved. Where a grade is not available, the nearest available grade is substituted. Standards offering only Not Achieved or Achieved need four samples. Where too few students were assessed, all student work goes in, still capped at two at Not Achieved.

Moderators report on whether the school’s judgements are consistent with the standard, and a school that disagrees can query or appeal the decision inside the same application. Schools also receive an annual summary of overall consistency through their School Relationship Manager. Treat that summary as departmental data rather than administrivia, because it is the clearest external read on whether your marking sits at the national standard.

What does a Managing National Assessment review look at?

At least once every four years, NZQA reviews how a school manages assessment for national qualifications. The review covers four areas of practice: moderation, assessment practice, data and communication. Evidence comes from the school’s self-evaluation, documentation of assessment policy, interviews with school managers, teachers and students, and audits of the school’s systems.

The resulting report is published online, and where significant issues are found NZQA works with the school on an action plan to resolve them. The practical implication for a head of department is that the review reads your systems and your documentation, not your intentions. Verification that happened in a corridor conversation and was never recorded counts for very little when someone asks to see the evidence.

Where do AI marking tools sit in this?

Ministry of Education guidance on marking work with AI tools, published in October 2025 with NZQA, is principles-based. Teachers remain fully responsible for marking decisions. AI must support, not replace, teachers’ professional judgement. Human oversight is essential, teachers must review and validate any AI-generated feedback or marks, and students should be informed when AI is used for marking.

On NCEA the guidance is direct: using AI to mark NCEA internal assessments is discouraged, and should be limited to supporting the assessor’s own final judgements. The reasoning is that each judgement is a professional claim with legal standing, because it can contribute to the award of a listed qualification, and that responsibility sits with the assessor. The same guidance points to formative work, assessment for learning, as the better fit, alongside warnings about inconsistency between tools and the potential for bias.

Privacy is the other constraint, and it is easy to breach without noticing. Putting student work into a tool that is not information protected may breach New Zealand privacy law, and many models reuse the prompts they are given for further training. Read the terms before a class set goes anywhere near a tool, and check the position against your school’s AI policy, which has to cover acceptable use of AI as part of your authenticity policy.

So a platform offering expert essay marking is best treated as a drafting assistant on formative work rather than as a source of NCEA grades. Jeddle drafts feedback against criteria the teacher supplies and leaves every grade with the teacher, which is the arrangement current guidance supports. Whatever a department chooses, the obligations above do not move: the task still needs a critique before use, the grades still need verification by a subject specialist, and the school still has to explain how each judgement was reached.