Services — Test audit
An independent second pair of eyes on your test
You send a finished test, exercise set, or answer key. We verify it against seven checks and return a findings report, a corrected version of the material, and a certificate with a public verification URL.
The seven checks
Every finding is filed under one category, numbered as below. Basic audits cover checks 1, 2, and 6; full audits cover all seven and include the corrected version.
- 1
Answer key correctness
We solve every item ourselves before looking at your key, then compare. Key errors survive surprisingly long in tests that have been used for years — nobody re-checks an answer key that 'has always worked'.
- 2
Answer uniqueness
The single most common flaw in amateur tests: the intended answer is right, but a second option is also defensible. These items are unmarkable and open to appeal, and they are invisible to the person who wrote them.
- 3
Distractor quality
A reading item that can be answered from world knowledge, or by matching a word in the option to a word in the text, does not test reading. We flag items that are guessable without the text, and distractors nobody would ever pick.
- 4
CEFR level calibration
A test labelled B1 that requires C1 vocabulary in its gaps measures the wrong thing. We check vocabulary and grammar against the declared CEFR level, referenced to public vocabulary profiles and the CEFR descriptors.
- 5
Format conformance
If the material declares an exam-part format, we verify it against that part's public specification: text length, number of gaps or questions, task type, instruction wording. A 'Part 2 open cloze' with nine gaps instead of eight is not that task.
- 6
Language proofreading
Grammar, punctuation, spelling, unnatural collocations, and US/UK consistency across the whole paper — instructions and rubrics included, which is where most slips hide.
- 7
Cultural appropriateness & factual accuracy
Texts that state things that are no longer true, stereotypes that don't belong in a classroom, and topics that public exam bodies themselves avoid. Factual errors in a reading text undermine the items built on it.
What a flawed item looks like
A real pattern we see weekly, reconstructed from an anonymised test. The writer knows the intended answer, so the second defensible option is invisible to them.
As submitted
High severityChoose the correct option (B1 grammar test, item 12):
If I ______ more time, I would learn Italian.
A haveB had ✓C would haveD was having
Findings. Option A also produces a correct sentence — a first conditional with a present-tense condition (If I have more time, I would…) is marginal, but If I have more time, I will learn Italian shows the real problem: the stem does not force the second conditional unless the main clause is read first, and both A and B are accepted by many markers. The item also tests a structure listed at A2/B1 boundary while the test is declared B1 — acceptable, but flagged for the calibration record.
As returned
ResolvedCorrected item:
If I ______ more time, I would learn Italian — but I work sixty-hour weeks.
A will haveB had ✓C would haveD was having
Why it works. The added context rules out the open condition, and distractor A is changed so that no reading of the stem accepts it. One defensible answer remains; the item now measures what it claims to.
The full sample report shows seven findings across all categories — read it here.
Scope and limits
We audit finished material: tests, exercise sets, answer keys, placement tests, and mock papers, at levels A2–C2. We accept Word, PDF, and spreadsheet formats; scanned paper is fine if legible.
An audit is a review of the material, not of your teaching or your results. We do not predict pass rates, and the certificate attests that a review was carried out — it is not an accreditation or an endorsement by any examination board.
Where your material references a public exam format, we check it against that format descriptively. We are independent of all exam boards.
Have a test that needs checking?
Send us the material and the deadline. We reply with a fixed quote within one working day.