An IQA sampling plan is not a form to fill in — it is a documented, risk-based decision about where quality can quietly slip. Most centres treat it as a compliance ritual (a 10 % blanket, back-loaded to the end of the year, only summative decisions), and the external quality assurer sees the same pattern in every centre they visit. The plans that actually protect a centre look nothing like that.
This guide sets out what a sampling plan is for, how to build one using the widely used CADET model, what the plan should contain, what makes an EQA relax at your centre visit versus reach for more evidence, and where sampling plans routinely fail.
What the sampling plan is actually for
Internal quality assurance exists so learners, employers, awarding organisations and the wider public can trust that assessment decisions are safe, fair and consistent. The IQA cannot possibly re-check every piece of evidence — that would just duplicate the assessor’s work. Instead the IQA selects a representative sample, uses it to form a reliable picture of assessment quality across the centre, and evidences that picture in a plan and its accompanying records.
Two audiences read the sampling plan and its outputs:
- The EQA — external quality assurer from the awarding organisation. They are checking whether the centre’s IQA is genuinely operating, not just documented.
- Your own assessors — the plan should tell them, in advance, that their work will be sampled and roughly how often, so nothing about IQA is a surprise.
A sampling plan that only exists to satisfy the first audience gets found out fast.
The CADET model — what to sample against
CADET is the standard framework for building a sampling plan in UK vocational assessment. It captures the five dimensions your sample must cover to be representative:
- C — Candidate — a spread of learners across ability, background and stage of programme
- A — Assessor — every assessor on the qualification, weighted toward those newest or riskiest
- D — Delivery — every location, cohort and mode (workplace, classroom, remote)
- E — Evidence — every type of evidence used (observation, witness testimony, professional discussion, written work, product evidence, RPL)
- T — Type — every unit and assessment method on the qualification
A common failure is to sample enough volume but miss a whole dimension — for example, sampling 15 % of decisions but never once looking at witness testimony, or sampling every assessor except the one who is also the manager. CADET is a check on that: run through the five letters and you will find the gap.
Risk-based sampling — why percentages are the wrong starting point
The “10 % rule” is a folk custom, not an awarding organisation requirement. What awarding organisations actually require is that the sample size is proportionate to risk. Where the risk is high, sample more. Where the risk is low, sample less. The plan should evidence how that judgement was made.
Factors that push risk (and sample size) up:
- New assessor — first three months of practice, or first cohort on a new qualification
- Assessor with previous IQA concerns — action points from a prior sample not yet resolved
- New or changed qualification — first cohort where centre practice has not yet settled
- High-stakes evidence — RPL, professional discussion, or witness testimony where authenticity is harder to prove
- Learner-appeal history — any assessor or cohort with a live or recent appeal
- EQA action point outstanding — the specific area the EQA flagged last visit
Factors that reduce risk:
- Experienced assessor with a clean IQA history over multiple cohorts
- Evidence type with strong intrinsic authenticity (observed practical assessment)
- Standardised unit that has run cleanly for years
A defensible plan shows the risk rating for each assessor and unit, and the sample size flows from that rating — not from a flat percentage across the board.
Sample across the whole assessment cycle, not just at the end
The single biggest failure in UK IQA sampling plans is back-loading — sampling almost entirely once assessment is finished. By then, no early intervention is possible. A learner has been through weeks or months of assessment before anyone checks the assessor’s judgement.
A good plan samples at three stages:
- Interim (formative) — early in the learner journey. Catches assessor drift while there is still time to fix it and re-cover any unit that was misjudged.
- Ongoing — through the middle of the programme, confirming the interim finding held.
- Summative — before the final claim to the awarding organisation. Confirms the decision is safe to release.
Interim sampling should sit at roughly a third of the total. If yours is 5 %, the plan is back-loaded.
What a sampling plan document contains
There is no prescribed format — awarding organisations, and the widely used templates from OCN London, Focus Awards, Gateway Qualifications and ASDAN, differ in layout. But every workable plan contains:
| Field | Purpose |
|---|---|
| Qualification and unit(s) | Anchors the plan to a specific programme |
| Assessor | Named; ties sampling activity to a specific person |
| Cohort and delivery site | Ensures every cohort and location is represented |
| Risk rating (low / medium / high) with rationale | Shows the judgement — the rationale is more important than the rating |
| Sample size and % target | Flows from the risk rating, not the other way around |
| Evidence types to be sampled | Forces you to pick up witness testimony and RPL, not just the easy stuff |
| Stage of assessment cycle | Interim / ongoing / summative |
| Planned sample dates | Spread across the cycle, not clustered at the end |
| Actual sample dates | Filled in as the plan runs; gaps between planned and actual are themselves a signal |
| Findings summary | Cross-reference to the IQA report for each sample |
| Action points and closure dates | What was found, what was done about it |
Most awarding organisations offer their own template. Where they do, use it — an EQA reading their own centre’s own template will absorb it faster than a bespoke one.
Six sampling-plan failures EQAs flag most often
- The 10 % flat rate. Ignores risk, gives the impression the IQA is not thinking. Cite risk rating in every row.
- Back-loading. Almost all sampling happens after summative decisions. No opportunity to catch drift early.
- One-evidence-type sampling. Every sample is observation records or every sample is written work. Authenticity in the untouched evidence types goes unchecked.
- Assessor omission. The team lead, senior assessor or subject-matter expert is never sampled. If they are the strongest assessor, their samples become your training exemplars — do not skip them.
- Planned vs actual mismatch never explained. A row where planned was March and actual was June, with no note, tells the EQA the plan is a fiction.
- No cross-reference to the IQA report. The sampling plan says “sampled — no issues” with no linked report. That is not sampling; that is a claim of sampling.
What a defensible plan looks like from the outside
An EQA opens the plan and sees, in the first minute: every assessor listed, every unit covered, risk ratings that vary with a rationale, sampling spread across the year, multiple evidence types, and every planned row cross-referenced to a real IQA report. They do not need to ask questions to understand the centre’s sampling practice — the plan explains itself.
An EQA opens the failing plan and sees: 10 % in every row, all sampling in June, three of the eight assessors never sampled, only observation records in the evidence column, and “N/A” in the report reference. They then ask questions, and the visit gets longer.
Further reading
- Gateway Qualifications — Internal Quality Assurance guidance
- OCN London — IQA templates including sampling plan and tracking sheet
- Ann Gravells — IQA sampling reference
CPD training on IQA sampling
Our Internal Quality Assurance: Sampling Strategies and Monitoring course covers the purpose of IQA, how to design a sampling strategy using the CADET model, how to rate risk, and how to sample across the whole assessment cycle — with recording and feedback that stands up to EQA scrutiny.
Related: Standardisation and Moderation for Internal Quality Assurers · Standardisation and Moderation in Vocational Assessment · Providing Effective Feedback to Learners in Assessment
£9 per course. CPD accredited, 100 % online, verifiable e-certificate on completion. Pay for 2, get 3 — any 3 courses for £18.
Please note: this is an awareness-level CPD course. It develops your understanding and does not confer assessor or quality-assurer status; it is not a TAQA or CAVA unit or any Ofqual-regulated qualification. Where you need to be formally recognised as an internal quality assurer on a regulated qualification, the Level 4 Award in the Internal Quality Assurance of Assessment Processes and Practice from City & Guilds, Pearson or another Ofqual-recognised awarding organisation is the route.