Find out what the evidence says about your test
Validation research shows how reliable a test is and which interpretations of its scores the evidence supports, for a stated use, group and language.
Continuous recalculation on the platform is being built up per test as administrations come in.
What a validation study examines
Which analyses fit depends on the test and the use. These are the questions we usually start with.
Internal consistency
Do the items of a scale measure the same thing? We report Cronbach's α and McDonald's ω, which is more suitable when items differ in how strongly they reflect the construct.
Test–retest reliability
Are scores stable over time? Relevant for traits expected to be stable, with an interval that fits the construct and the intended use.
Content validity
Do the items cover the construct as defined, without parts that belong elsewhere? Assessed by expert review against the construct definition.
Factor structure
Does the data show the scales the test claims? Exploratory and confirmatory factor analysis test whether the structure holds, per language version.
Convergent and discriminant validity
Does the test relate to measures of similar constructs, and less to measures of different ones? This places it in a wider network of evidence.
Criterion validity
Do scores relate to outcomes that matter, such as performance ratings, training results or retention? This needs a relevant, well-measured criterion.
Three ways to work with the Lab
Prices for each service are on the Psychometric Lab page. The intake shows which one fits.
Quick scan
We analyse an existing test and the data you already have, and tell you what the evidence supports and where the gaps are.
Validation study
Reliability, construct validity, item analysis and norms for a defined use, based on a study design agreed in advance.
Full validation
Research design, norming, fairness analysis and a technical manual you can publish, for tests that need a complete evidence base.
From study to psychometric label
On Match.hr, the results of a study appear on the test's psychometric label, so users can see what the evidence covers.
- Reliability coefficients with the sample they are based on
- The intended use the evidence applies to
- Which language versions have been analysed
- The review level, from Listed to External review
- The date of the last analysis
Our review criteria follow the structure of the COTAN and EFPA test review models. Match.hr is not affiliated with or endorsed by either organisation.
Numerical reasoning
Example- Intended use
- Selection, development
- Reliability (ω)
- 0.89
- Norm sample
- 12,482
- Languages validated
- EN, NL, DE
- Fairness analysis
- Completed
- Last analysis
- October 2026
What validation can and cannot show
A validation study does not certify a test. It shows what the evidence supports for a stated use, group and language, and it may show that it supports less than hoped. Coefficients from small samples are unstable, and criterion studies usually need several hundred participants with a relevant outcome. Evidence for one language version, role or use does not carry over automatically to another.
Frequently asked questions
What reliability is good enough?
It depends on the use. Decisions about individuals, such as selection, ask for higher reliability than research at group level. Review models give rough thresholds, but we report the coefficient with its sample and confidence interval, and the standard error of measurement, instead of a simple pass or fail.
Do we need our own data?
Not for a first look. A quick scan works with the data and documentation you already have. If there is little or no data, a validation study starts with a research design and data collection, which can run on Match.hr.
What does Validated mean on Match.hr?
It is a review level of the Lab: sufficient evidence of reliability and validity for the stated use. It is not an official review. Our criteria follow the structure of the COTAN and EFPA test review models, and we follow the principles of ISO 10667 for assessment in work settings. Match.hr is not affiliated with or endorsed by these organisations, and not certified against that standard.
Is a test validated once and for all?
No. Validity is about the interpretation of scores for a particular use. When the use, target group, language or test version changes, the evidence has to be looked at again. On Match.hr, item statistics and norms are recalculated as administrations grow, so changes become visible sooner.
Related pages
Test your first candidate today
Free account, no credit card, results in minutes.