Your exams, moderated before they go to print.
Commission us either way: we independently moderate your teachers' own drafts before administration, or write unseen base papers and parallel/resit versions from scratch against your specification. Both paths give you item-to-outcome and CEFR-descriptor mapping, structural item checks, and a human-signed moderation record.
Scoped to your levels, your test calendar and your delivery platform. No student data accepted. No publisher PDFs required.
Three fault classes that survive a proofread.
These are categories our record is built to surface. They are not a claim about how often they occur.
Key-position imbalance
Correct answers drift toward one option slot. Standard item-writing guidance is to vary the location of the right answer; balanced keys also make randomised item banks safer for automated scoring.
Construct & alignment drift
An item that tests the wrong sub-skill, sits off the target CEFR descriptor, or exceeds the specified text length or audio rate. Fair on paper, unfair to the candidate.
Appeal-risk ambiguity
Two defensible keys, or a distractor that could arguably be correct. One challenge from a candidate and the sitting is contested.
The moderation record, not a brochure.
Two outputs from the process, using synthetic A2/B1 items. Click between them.
Moderation record — 12-item sample
| Stage | Finding | Status |
|---|---|---|
| Test specification present | Spec and blueprint supplied; version equivalence defined | Pass |
| Item→outcome / CEFR mapping | 12/12 items mapped to descriptor and cognitive process | Pass |
| Key-position balance | Key positions clustered toward one slot in this synthetic set | Review |
| Stem / option homogeneity | 2 items: keyed option markedly longer than distractors | Review |
| Distractor plausibility | 1 item: distractor could arguably be correct | Fix |
| Construct-irrelevant variance | 1 item: input text above the specified word count | Fix |
| Negative / absolute wording | None found | Pass |
| Transport integrity (CSV / XML / QTI) | Verified against target platform version | Pass |
Key distribution — synthetic set
Illustrative synthetic paper, shown only to demonstrate the check. Recommended practice (Haladyna, Downing & Rodriguez, 2002, guideline 20) is to vary the location of the right answer across a paper.
Item map (excerpt)
The Night Bus Library. A short A2/B1 reading text used only to illustrate mapping.
| Item | Sub-skill | Descriptor | Process | |
|---|---|---|---|---|
| R1 | Reading · gist | B1 understanding main points | Careful local reading | ok |
| R2 | Reading · detail | B1 identifying specific information | Scanning | ok |
| R7 | Reading · inference | B1 inferring meaning | Inference | amb. |
| G2 | Use of English | A2–B1 grammatical accuracy | Form recognition | ok |
| V2 | Lexical | B1 lexical range | Lexical inference | ok |
Descriptor = CEFR Companion Volume scale and level. Process = the mental process the item intends to elicit (ALTE Manual for Language Test Development and Examining, §3.1.2.1).
Parallel versions — illustrative
| Item | V1 | V2 | V3 | V4 |
|---|
Illustrative only. Versions are one option: we also build rolling LMS item pools with randomised delivery if that fits your calendar.
Whether your centre needs parallel paper versions, a main sitting plus a resit/medical paper, or an LMS item pool depends on your intake model and your awarding rules — we scope to yours.
Vendor-neutral exports
Sample files, delivered as in the pack. Every export is verified against your platform version before hand-over.
- Answer key — generic CSV / OMRCSV
- Moodle question bankXML
- Item-to-descriptor mapCSV
- Moderation recordHTML
Also supplied: QTI 2.1 (Canvas / Blackboard Ultra / Inspera) and QTI 2.2 (on request), DOCX / PDF print masters, and a multi-speaker listening file. Platform notes cover Inspera, WISEflow, Exam.net, Respondus, Questionmark and lock-down browsers such as Safe Exam Browser.
One free diagnostic. Scoped to your assessment volume.
We price by assessment volume you control — not by a fixed level or form count — so the same service fits a 3, 5 or 7-band centre.
Two ways to work with us, priced by assessment volume: moderation-only (independent audit of papers your own teachers write) or full unseen build + moderation (we write the base paper and its parallel/resit version). Same moderation record either way.
- Item-to-CEFR / outcome map
- Sections A–G pre-administration moderation findings
- Delivered within 24 hours
- Moderation-only (from £395): external pre-administration audit of your in-house draft paper(s), item-to-outcome & CEFR Companion Volume mapping, structural fix notes (Haladyna et al.), and a human-signed Sections A–G moderation record
- Unseen build + moderation (from £650): unseen base paper + parallel/resit version written to your specification, moderated, signed off, and exported to CSV / Moodle XML / QTI 2.1 / DOCX
- Moderation-only (from £1,250/yr): scheduled termly or rolling-intake external moderation across your CEFR bands, giving continuous inspection-ready assessment records
- Full build + moderation (from £1,950/yr): the same programme with full unseen paper + resit engineering across the year
Optional add-ons, quoted separately: Writing/Speaking rubric & standardisation moderation; additional parallel/resit versions; migration of an existing bank to QTI. If you would rather build capability in-house, we can run a standardisation or CPD session on request — it is not bundled with the assessment work.
Built to pass a procurement and quality review.
Frameworks, not endorsements
Our record is designed to help you evidence common quality expectations (e.g. EALTA & ALTE guidance; BALEAP BAS criterion II.6–II.7; EAQUALS assessment category 7.1–7.5; British Council criterion T16 and placement linkage). We are not accredited, endorsed or certified by any of these bodies; accreditation is theirs to award to your centre, not to us.
Copyright-clean by design
You send a syllabus matrix, unit list and test specification — never a publisher PDF. We write original items against your specification.
Zero student PII
No names, IDs, marks or scripts are accepted. If personal data is found in a submission, the job is returned.
Independent human moderation
Every item, key and rationale is reviewed and signed off by a moderator before delivery. The audit trail is yours to keep.
Leadership. Led by a serving Head of an English for Academic Purposes programme, with experience across university pathway, foundation and high-volume computer-based assessment operations.
What a Director of Studies usually asks.
What do you need from us to start?
Your test specification or blueprint, CEFR levels, unit themes, target language, item counts and delivery platform — and, for the diagnostic, one current paper (a teacher-written draft you want moderated, or a template you want us to build against).
Who writes and checks the items?
Items are authored against your specification, then every item, key and rationale is reviewed and signed off by a human moderator before anything is delivered.
Do you accept student data?
No. We never accept names, IDs, marks or scripts. The work is on the assessment, not the people sitting it.
What about copyright?
We do not use publisher PDFs. All items are original and written to your specification, which keeps the output clean for internal use.
How is the work scoped?
By assessment volume, in one of two modes. Moderation-only: we audit, map and sign off papers your own teachers write. Unseen build + moderation: we write the base paper and its parallel/resit version to your specification, then moderate and sign it off. Either mode can be a single cycle or an annual programme, so it fits centres running different numbers of CEFR levels and intake models.
We already write our own papers. Can you just moderate them?
Yes. Many centres only want ongoing external pre-administration moderation: an independent audit of teacher-written drafts, item-to-outcome and CEFR Companion Volume mapping, structural fix notes (Haladyna et al.), and a human-signed Sections A–G moderation record. That is the moderation-only line, from £395 a cycle. You keep authoring; we keep it defensible.
How fast is the free diagnostic?
Within 24 hours of receiving one sample paper and your specification, you get the item-to-descriptor map and a summary of structural and alignment findings.
What formats do you export?
Generic CSV answer keys / OMR, Moodle XML, QTI 2.1 (Canvas, Blackboard Ultra, Inspera) and QTI 2.2 on request, plus DOCX/PDF print masters and the signed moderation record. We verify the import on your platform version.
Do you assess Writing and Speaking?
Objective papers are the core. Writing and speaking rubric alignment and standardisation can be added, scoped to your task types and marking scheme.
Send one paper. See the moderation record for free.
One sample exam and your specification. Within 24 hours you get the item map and the structural and alignment findings. No obligation, no student data.
Request Free 24-Hour Diagnostic