Norm groups that fit the people you assess
A raw score only means something next to a comparison group. The Lab builds norm groups per country, language and role, and updates them as samples grow.
Norm updates on the platform are being built up per test and language as administrations come in.
What a norm score means
A position, not a verdict
A norm score says where someone stands compared with a particular group. It does not say whether that person is suitable for a role.
Percentiles, stanines, T-scores
Different scales for the same position. A percentile of 70 means 70 per cent of the norm group scored lower; stanines and T-scores group or standardise that position.
A range, not a point
Every score has a measurement error. The report shows a confidence range, and small differences between people often fall within it.
What a good norm group needs
A norm group is only useful when it resembles the people you compare with it, and when you know who is in it.
- A clear description: country, language, role or education level, period of data collection
- Enough people overall and in every subgroup that gets its own norms
- A composition that reflects the population the test is used for
- The same test version, language and administration conditions as in practice
- Cleaned data: incomplete, careless or duplicate responses removed by documented rules
- A date, so users can see how recent the norms are
How norms are built and maintained
Define the reference group
Agree whom the scores should be compared with: applicants for a type of role, employees in a sector, or the working population of a country.
Collect and clean the data
Gather administrations under comparable conditions and apply documented rules for excluding invalid responses.
Compute the norm tables
Convert raw scores to norm scores, check the distribution for ceiling and floor effects, and smooth where samples are thin at the extremes.
Monitor and update
Recalculate as data grows, compare with earlier norms and record each update, so a report can be traced to the norms it used.
Norms on the psychometric label
On Match.hr, the label of each test shows how large its norm samples are and when they were last calculated.
- Norm sample size per country and language
- The norm group named in every report
- The date of the last norm calculation
- Recalculation on an anonymised dataset as administrations grow
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
Sample size and its limits
There is no single minimum. Several hundred people per norm group is a common starting point; finer norms, separate subgroups and reliable extreme percentiles need more. Small norm groups give unstable norm scores, especially at the top and bottom. Norms from one country or language are not reused for another, because a translation can change how a test works.
Frequently asked questions
Should we use a general or a specific norm group?
Use the group you actually want to compare with. For selection that is usually applicants or employees at a similar level, not the general population. A very specific group fits better but is often smaller, so the norms are less precise. The intake helps you weigh the two.
What happens to scores when norms are updated?
The same raw score can lead to a slightly different norm score. That is why the report names the norm group and its date. Compare candidates in one procedure against the same norms, and keep earlier reports with the norms they were based on.
Can we have a norm group for our own organisation?
On request, once there is enough data. An internal norm compares people with your own employees or applicants, which can be useful for development. For selection, check that the group is large enough and that it does not carry over earlier selection choices into new decisions.
Do you make separate norms for men and women?
Not by default. In selection, separate norms per gender or other protected characteristic can amount to unequal treatment, and they hide differences that should be examined. We look into such differences in a fairness analysis, where lawful, and report them at group level.
Related pages
Test your first candidate today
Free account, no credit card, results in minutes.