Fixed-price product

Measurement Quality Audit

A fixed-scope, fixed-price psychometric diagnostic of a survey, scale, index or test you already run. Two tiers, a written verdict signed by a named psychometrician, and nothing to scope before you start.

Ask about an auditRead a sample report

Why audit

Every instrument makes a claim.

Its score means what it says, and means the same thing for everyone it is used on. Most instruments were never tested against that claim, and a translation, a new site or a new population quietly breaks it. The audit tests the claim and tells you, in writing, what your scores can and cannot support.

Two tiers

Choose by what you have, not by what you can afford to scope.

The desk review needs only the questionnaire. The full audit adds your pilot or field data. Both are fixed in scope, price and turnaround, and both end in a verdict page you can hand to a board or a funder.

Tier 1. Instrument Review

A structured expert review of the instrument as it stands. No data needed.

You send

The instrument as administered: items, instructions, response options, scoring key
Your construct definition, or the document that stands in for one
Any translated versions, with whatever translation record exists

You receive

A scored item table with the reason for every call: keep, revise or cut
A construct-coverage, response-scale and scoring verdict
A translation equivalence-risk table for each language version
A prioritised fix-list in three severity bands
A signed verdict page and a thirty-minute findings call
Turnaround
Ten working days from receipt of the instrument.
Scope ceiling
One instrument, up to 60 items or eight subscales, up to three language versions.
Price
1,800 EUR
Ask about an audit

Tier 2. Full Psychometric Audit

Tier 1 plus a quantitative pass on your own pilot or field data.

You send

Everything in Tier 1
One data file with one row per respondent, and a codebook
A grouping variable, where you want groups or languages compared

You receive

Item analysis, reliability (alpha and omega) and factor structure against the claimed model
The measurement-invariance cascade across groups or languages, with partial invariance sought where a level fails
A technical report to APA seventh edition, with every table ready to paste into your own reporting
The scripts that reproduce every number
The fix-list, a signed verdict page and a sixty-minute findings call
Turnaround
Four weeks from receipt of clean data.
Scope ceiling
Up to 60 items, 5,000 respondents and four groups on one grouping variable, one measurement occasion.
Price
5,900 EUR
Ask about an audit

Prices exclude VAT.

The boundary

Where an audit stops and a study begins.

The ceiling is the product. Beyond it, the work is a validation study and is quoted as one. Outside both tiers, and quoted separately:

New data collection or a pilot
Norms, reference distributions, cut scores or standard setting
Criterion or predictive validity against an external outcome
Longitudinal or multi-occasion invariance
Rewriting the instrument or authoring new items

The fix-list in either tier is written so that you can commission exactly the follow-on you need, or none.

Evidence

More than one expert's opinion.

The best current evidence on structured quality review is a 2026 preprint, not yet peer reviewed, in which minimally trained raters assessed the methodological rigour of 52 and then 110 psychology papers against written criteria (Etzel et al., 2026, PsyArXiv 4w7rb). Overall rigour scores agreed well across raters, and typical recent papers met under a tenth of the criteria. Two limits, stated plainly: the paper rates papers, not instruments, and agreement was strong for checkable criteria and weak for judgement calls. The audit is built to that finding. Every call in the report is tied to a written criterion and a quoted item so a second psychometrician can check it, and AdriaMont is running its own two-auditor agreement study on the first audits.

Public worked example

Ranks without resolution.

A measurement audit of a multilingual language-model benchmark, asking how much of its language ordering is actually estimable. The data, code and every results table are open.

Open the deposit (Zenodo)

Sample reports

Tier 1 sample: desk review of the Rosenberg Self-Esteem Scale (PDF)Tier 2 sample: technical appendix, three-language invariance run (PDF)

Both samples are demonstrations on public-domain or synthetic material. A client receives the same documents on their own instrument and data.

Start an audit

Send the instrument. Get a date.

Tell us which tier, attach or describe the instrument, and we reply within two working days with a start date and the written scope.

Or write directly to

milos@adriamont.me
Questionnaire Pre-screenInstrument DevelopmentResearch, Evidence & EvaluationView all services