ESOL and LLN Assessment Design for NZQA: Fix the Standard
One unit standard, many readers. NZQA already pegs NCEA Literacy and Numeracy unit standard 32403 to three reading scales at once. The lesson for PTEs is to hold the standard constant, vary only the language load, and build assessments that can be re-mapped when skill standards replace unit standards.
Why does one unit standard need so many reading levels?
Because your learners don't arrive at one level. A single cohort can include a confident native reader, a learner on the Adult Learning Progressions, and an ESOL learner who knows the trade better than the question paper lets them show.
"Ten reading levels" is our shorthand, not a sourced statistic. The point is real, though. If the assessment's language is harder than the standard's own demand, you are testing English, not competence.
That is the ESOL and LLN assessment bottleneck. Writing one version per profile doesn't scale, and writing one version for everyone fails the learners at the edges.
What does NZQA's own practice show?
New Zealand already treats "one standard, many scales" as normal. NZQA's 2026 assessment specifications for NCEA Literacy and Numeracy unit standard 32403 benchmark the reading standard against:
- the NZ Curriculum, at upper Level 4 to lower Level 5
- Adult Literacy Learning Progressions, Step 4
- English Language Learning Progressions, upper Stage 3 to lower Stage 4
The same common assessment activity runs online in set windows (18–29 May and 7–18 September 2026). One task, one standard, and very different English-language profiles sitting it.
That is a useful design principle. The standard is the fixed point. The language scales are how you check the task is fair.
Where does the bottleneck actually show up?
In the achievement data. In Event 1 2026, Reading had 59,003 participants and 60.5% achieved. Numeracy had 61,846 participants and 51.7% achieved.
The gap between groups is starker. Reading achievement was 73.0% for the "Fewer" socioeconomic-barrier group against 36.9% for the "More" group (NZQA, 2026 additional data).
Two caveats matter. This is NCEA data, not PTE or polytechnic data. And the excerpt we reviewed didn't break results down by ethnicity or ESOL status. We found no equivalent tertiary-level ESOL and LLN dataset, so treat these figures as a signal, not proof about your learners.
Demand for English-language support is also rising, though again the evidence is school-sector. A record 90,770 students received ESOL funding (Newstalk ZB, June 2026). Budget 2026 added $10m, taking the annual budget to $73m, without raising the funding rate.
What is shifting around LNAAT and the standards themselves?
Two things, and they land close together.
LNAAT stewardship is consolidating at TEC. The Literacy and Numeracy for Adults Assessment Tool was previously managed by NZCER and is now supported by TEC. Its web address, functionality and use are unchanged. A TEC notice says the Adult Learning Progressions and teaching and learning resources hosted by Ako Aotearoa move to the TEC website on 20 October 2026. Check your internal references and learner materials for old links.
TEC's LNAAT guidelines set out usage expectations by fund. Its Starting Points guide covers learners not yet ready for progressions-based assessment, which is the route for very low-level learners, many of them ESOL.
Standard-setting is moving. Workforce Development Councils were disestablished on 31 December 2025, and eight Industry Skills Boards now provide standard-setting. A February 2026 Ministry of Education Cabinet paper says the Boards will handle standards setting, qualifications development, endorsing programmes and moderating assessments. Skill standards are set to replace unit standards.
What we couldn't find is guidance on how skill standards will treat foundation skills. Verify the legislation dates and any 2027 or 2028 assessment deadlines against TEC primary sources before you plan around them.
How do you design for ESOL and LLN without rewriting everything?
Treat the standard as the constant and the language as the variable. A workable sequence:
- Pin the evidence requirement. Write down exactly what the standard asks the learner to do, in plain words, before you draft a single question.
- Separate the skill from the language. For each task, ask what reading demand it adds beyond what the standard itself requires. Cut anything that isn't needed.
- Set a language-load range. Use the Adult Learning Progressions and, for ESOL, the English Language Learning Progressions as your yardsticks, as NZQA does for 32403.
- Vary context, not outcome. Swap the industry scenario, sentence length, scaffolding or response mode (oral, visual, written) while the assessment decision stays identical.
- Map it back. Keep a clear trail from each task to the standard clause, so re-mapping to a skill standard is an edit, not a rebuild.
Step 5 is where most providers lose time. Hand-built mapping in spreadsheets is slow and brittle. Platforms such as VETos generate assessments in an industry and learner context and map them to the standard automatically. Whatever you use, the principle stands: mapping should be a living artefact, not an afterthought.
What should you do before skill standards land?
You don't need the final skill standards to start. You can:
- audit which assessments carry unnecessary language load
- tag each task to its standard clause so the mapping is portable
- fix any references to the old Ako Aotearoa-hosted resources
- record how you place learners (LNAAT or Starting Points) alongside your assessment variants
The providers who do this now will treat the reform as a re-labelling job. The rest will treat it as a rewrite.
Key takeaways
- NZQA unit standard 32403 is benchmarked against the NZ Curriculum, Adult Literacy Learning Progressions and English Language Learning Progressions at once.
- The right design goal is a fixed standard with a flexible language load, not separate assessments per learner.
- LNAAT support has moved to TEC, and Adult Learning Progressions resources move to the TEC website on 20 October 2026.
- Industry Skills Boards now handle standard-setting, and skill standards are set to replace unit standards, so portable mapping matters.
- Tertiary-level ESOL and LLN data, and guidance on foundation skills in skill standards, are still gaps. Verify dates with TEC before committing.
Our take
Most ESOL and LLN "support" is bolted on after the assessment is written, which is why it's slow and inconsistent. NZQA's own approach to 32403 shows the better order: fix the standard, then check the language demand against known scales. We think providers who build that discipline into assessment design now will absorb the move to skill standards far more cheaply than those who wait for the wording to be final.
FAQ
What is NCEA Literacy and Numeracy unit standard 32403?
It is the NCEA Literacy and Numeracy reading standard. NZQA's 2026 assessment specifications benchmark it against the NZ Curriculum (upper Level 4 to lower Level 5), Adult Literacy Learning Progressions Step 4, and English Language Learning Progressions upper Stage 3 to lower Stage 4.
Who manages the LNAAT now?
The LNAAT was previously managed by NZCER and is now supported by TEC. TEC states its web address, functionality and use are unchanged. Adult Learning Progressions and related resources hosted by Ako Aotearoa move to the TEC website on 20 October 2026.
Will skill standards change how ESOL and LLN assessments are mapped?
Probably, but it isn't settled. Skill standards are set to replace unit standards under the Industry Skills Boards, and we found no guidance yet on how they treat foundation skills. Keep your mapping portable and check TEC and the relevant Industry Skills Board for updates.
Is there PTE or polytechnic data on ESOL and LLN achievement?
We didn't find any in the sources reviewed. The available figures are NCEA data (for example Event 1 2026 Reading at 60.5% achieved) and school-sector ESOL funding numbers, so apply them to tertiary learners with care.