---
title: "Norm group"
description: "A norm group is the reference sample a raw score is compared against in order to produce a percentile or standard score. The same performance can be an unremarkable result against one norm group and an exceptional one against another, which makes the identity, size and date of the norm group part of the score itself."
canonical: https://www.assessall.com/guides/glossary/n/norm-group
updated: 2026-09-25
source: AssessAll
---

# Norm group

_Also known as: Normative sample, Comparison group, Norms._

A norm group is the reference sample a raw score is compared against in order to produce a percentile or standard score. The same performance can be an unremarkable result against one norm group and an exceptional one against another, which makes the identity, size and date of the norm group part of the score itself.

<!-- #the-most-important-line-in-a-score-report-and-usually-the-sm -->
## The most important line in a score report, and usually the smallest

A percentile is meaningless without knowing who it is a percentile of. 'Better than 85% of people' is a different claim from 'better than 85% of applicants to graduate schemes at large financial firms in 2019', and only the second is actually being made.

Ask four things of any published norm: who is in it, how many of them there are, when it was collected, and whether it is an applicant sample or an incumbent sample. Incumbent norms are drawn from people who already passed selection, so they are systematically stronger and will flatter nobody.

There is a fifth question worth asking of any battery that reports several scales on one scaling, and it is the one nobody asks: is the same scale value the same percentile on every scale? It usually is not, because a standard score that is a straight linear rescaling of the raw score keeps the shape of that scale's own distribution, and different scales are skewed differently. This is a solved problem in test publishing rather than an obscure one — the MMPI-2's uniform T scores were built precisely so that a value would mean the same percentile everywhere, described in the literature as "percentile-comparable, yet, unlike normalized T-scores, depart minimally from the familiar linear T-scores". If a report puts two scales side by side on the same axis, ask which of the three it is using.

<!-- #when-there-is-no-norm-group-at-all -->
## When there is no norm group at all

Sometimes the honest answer is that there isn't one yet, and a small sample is not a norm. A distribution built on a few dozen people gives percentiles that will move substantially when the next few dozen arrive — and "substantially" can be put in figures, which almost no published norm does.

The alternatives are legitimate and should be labelled as what they are: a criterion-referenced result reported against a defined standard rather than against other people, or a cohort comparison stated explicitly as a position within this particular applicant pool.

---

Source: https://www.assessall.com/guides/glossary/n/norm-group