ESOL Assessment Design NZQA: Closing the Validity Gap

New Zealand's own literacy and numeracy assessment tools carry an acknowledged validity gap for ESOL learners, particularly in numeracy, and the risk extends well beyond foundation-funded cohorts. The sector's established remedy — embedding literacy, numeracy and language support directly into assessment content rather than bolting it on afterwards — is widely recommended but rarely built at scale, leaving PTEs, ITOs and ITPs exposed on both equity and moderation grounds.

What does the research actually say about ESOL and NZ's assessment tools?

This isn't speculation. Government-commissioned research into adult English language learners found that while reading, writing and vocabulary results tracked English proficiency roughly as expected, numeracy results were variable. The researchers stated plainly that it remains unclear whether numeracy assessments can be used reliably with English language learners.

That's a direct admission, published through Education Counts and The Hub, that a core national assessment approach may not validly measure numeracy for a significant slice of NZ's learner population. For any provider assessing against unit standards where numeracy is embedded, that's not a footnote — it's a live compliance question.

Is this only a foundation-level problem?

No. TEC's 2024 Guidelines for using the Literacy and Numeracy for Adults Assessment Tool (LNAAT) confirm that literacy and numeracy gaps may present barriers to participating and succeeding in education and work, even for learners sitting above foundation level.

That matters because a lot of assessment design logic in the sector still treats LLN and ESOL support as something that applies to foundation-funded programmes only. TEC's own guidance says otherwise:

• Gaps can affect learners across the NZQF, not just entry-level cohorts.
• Barriers show up in workplace, classroom and distance delivery alike.
• Standard unit standard assessment, if generic, can miss these learners entirely.

Why does oral fluency make this harder to spot?

The Ministry of Education's January 2026 ESOL bulletin flags a specific trap: oral fluency can sometimes mask challenges with academic literacy. A learner who converses confidently in English may still struggle with the written, technical or numeracy demands of a unit standard assessment.

The Ministry's guidance is explicit that providers need a clear, consistent process to identify each student's starting point. That's an assessment design problem, not just a pastoral care one — and it's not one that a generic, off-the-shelf assessment template is built to solve at scale.

What can vocational providers learn from NCEA's co-requisite stumble?

The NCEA literacy and numeracy co-requisite is a schooling-sector reform, but NZCER's analysis of it is a useful cautionary case study for anyone building high-stakes standalone assessment. NZCER found pass rates as low as 56 percent on numeracy and writing tests in pilot data — a standalone, high-stakes digital testing model producing exactly the kind of gap the sector says it wants to avoid.

NZCER's recommendation wasn't more standalone testing. It was AI-assisted formative assessment, embedded in learning, with automated scoring and feedback built in as the learner progresses — not bolted on as a single high-stakes checkpoint at the end.

Embedding vs bolting on: the fix sector guidance already recommends

NZ tertiary guidance, including WelTec's professional development materials, defines embedding as weaving reading, writing, listening or maths instruction directly into the course content — the opposite of delivering it as a separate module.

Most PTE and ITO assessment development still runs the other way: a generic unit standard assessment gets built first, and LLN or ESOL support gets attached afterwards as a standalone add-on. That ordering is the structural problem.

What does this mean for PTEs, ITOs and ITPs building assessments?

The exposure here runs in two directions.

1. Equity risk — learners who need embedded, context-aware support are assessed against tools with an acknowledged validity gap, particularly in numeracy.
2. Compliance risk — providers who can't demonstrably show assessment coverage tailored to their actual learner cohort face harder moderation conversations with NZQA, not easier ones.

Building assessments that are mapped directly to the unit standard, generated in the context of the specific learner group — ESOL, LLN, workplace, distance — and supported by clear assessor guides isn't a nice-to-have layer on top of compliance. Based on TEC's own guidelines and the Ministry's bulletin, it's closer to what the standard already implies a provider should be doing.

Key takeaways

• Government-commissioned research admits numeracy results for ESOL learners are variable and it's unclear whether current tools measure numeracy reliably for this group.
• TEC's 2024 LNAAT guidelines confirm literacy and numeracy barriers extend beyond foundation-funded cohorts, across the NZQF.
• The Ministry of Education's January 2026 ESOL bulletin warns that oral fluency can mask academic literacy gaps, requiring a consistent process to identify each learner's starting point.
• NZCER's review of NCEA's co-requisite testing found pass rates as low as 56 percent and recommended embedded, AI-assisted formative assessment over standalone high-stakes tests.
• Sector guidance (e.g. WelTec) already defines the fix as embedding literacy and numeracy into content design — most PTE and ITO assessment development still does the opposite.

Our take

The uncomfortable part of this research isn't that ESOL learners struggle with generic assessment — it's that the sector's own regulators and researchers have said so in writing, and the standard response has still been manual patching rather than structural redesign. Embedding literacy, numeracy and learner context into assessment from the first draft, mapped straight back to the unit standard, isn't a compliance nicety. It's what TEC's guidelines and the Ministry's own bulletin are already pointing towards — providers who keep treating LLN and ESOL support as a bolt-on are building moderation risk into every unit standard they assess against.

FAQ

Does NZ have an official validity gap for ESOL learners in literacy and numeracy assessment? Yes. Government-commissioned research on adult English language learners found numeracy results were variable and stated it remains unclear whether numeracy assessments can be used reliably with this group, as published via Education Counts and The Hub.

Is this issue limited to foundation-funded ESOL learners? No. TEC's 2024 LNAAT guidelines confirm that literacy and numeracy gaps may present barriers to participating and succeeding in education and work even for learners above foundation level, meaning the risk extends across vocational programmes.

What does 'embedding' literacy and numeracy actually mean for assessment design? Sector guidance, including WelTec's professional development materials, defines embedding as weaving reading, writing, listening or maths instruction directly into course content — rather than delivering LLN or ESOL support as a separate, standalone module.

What can PTEs and ITOs learn from the NCEA co-requisite experience? NZCER's analysis of NCEA's co-requisite found standalone, high-stakes testing produced pass rates as low as 56 percent on numeracy and writing, and recommended embedded, AI-assisted formative assessment with automated feedback instead — a relevant caution for any provider relying on generic, standalone assessment tools.

Why does oral fluency matter for identifying assessment needs? The Ministry of Education's January 2026 ESOL bulletin warns that oral fluency can mask challenges with academic literacy, and stresses that providers need a clear, consistent process to identify each learner's actual starting point before assessment.