Significantly enhance the interpretation and
credibility of assessment results.
Benchmark group performance.
Login to the
NormMaker Norm Generator
effortlessly create norms for psychometric tests and instruments used in selection,
assessment, personal and group development. Full norm tables allow you to evaluate
performance using stanines, stens, percentile distribution, or standard scores. Quality
control of your instrument is provided for by descriptive statistics, reliability
coefficients and indications of the characteristics of the distribution. You can also
view your data by graphing the distribution profile, including the position of stanines.
NormMaker can use data in Ascii, DBase and even frequency table form to
allow users to enter test or questionnaire results easily or process existing scores
stored in almost any format.
Attractive and easily readable output for norm tables can be viewed on
screen, immediately printed, or saved to disk for future reference.
Reliability coefficients and other statistics for the test or
questionnaire are provided with each norm table.
Graph your data showing the overall distribution of scores, scores
distributed by stanine, or the top and bottom third of the distribution.
A comprehensive manual covering the use of NormMaker and an in-depth
discussion of the concept and use of norms is supplied with the program.
CONCEPT OF NORMSThe completion of a
test or questionnaire results in a raw score reflecting how the person has accomplished
the task. However, knowing a raw score does not tell a great deal about the person's
actual performance on the test or questionnaire. In order to evaluate performance, we need
a standard against which we can compare. In this way we are able to evaluate the person's
performance relative to a group who we are interested in. Norms provide us with such a
context. For example, if a person has obtained a raw score of 90 out of 100 for a test of
some kind, we may initially think this performance is very good. However, if we compare
this score to other people we are interested in and find that they score 95 or higher, it
means that the performance of the person we are looking at is in fact low relative to the
Further, using raw scores, we cannot identify how a person has performed across a range
of tests or questionnaires unless we can get a standard which can be applied to all the
tests. Using norms to evaluate performance, we may find that somebody performs strongly
using language skills, but is relatively weak in computational skills. Norms therefore
provide a frame of reference with which to view a candidate's performance.
Norms are constructed from a set of scores of candidates that are representative of the
population you are measuring or comparing against. A candidate's scores can then be
directly compared with the norm group, and the relative performance of the candidate can
be identified by how the candidate falls in the distribution of scores on the norm table.
In this way one can tell whether a candidate is anything from very poor or low, to
average, to exceptional on an ability or characteristic.
Lets have a look at some examples of how norms can be used:
i. When assessing a bursary candidate for his or her suitability for university, it is
useful to know how the person compares against students who have been successful at that
level. By comparing the performance of the candidate on a number of psychometric tests
against the norm group of successful students, you can evaluate the likely success of the
candidate in obtaining the qualification.
ii. In a selection exercise within an organisation you want an overall profile of how
good the candidates are on a range of measures. These may include psychometric tests and
questionnaires. Use of a norm table composed of all internal personnel provides a profile
of the performance of all candidates on the same criteria, and shows the relative
strengths and weaknesses of individuals as well as how they compare to the standard in the
iii. As part of a development scheme for personnel, a company may offer an
assertiveness course. Candidates complete a questionnaire on assertiveness and compare
their scores with those of a norm composed of a representative sample of the general
population. They can then see exactly how assertive they are relative to the next person.
iv. Use is made of an imported questionnaire as part of an organisation's management
training program. The questionnaire measures conflict styles. Using norms generated from
the scores of South African managers, the trainees can compare their own profiles with the
norm and begin to adapt their behaviour where necessary.
Measurements taken on large groups of people will generally create a distribution which
takes the approximate form of a normal bell shaped curve. For example, if we have a look
at the shoe size of the adult male population, it will fall in a range of scores from very
small to very large, with the majority of people somewhere in the middle.
The normal curve is characterised by a concentration of observations in the middle and
as the distance from the central point increases on either side, the incidence of people
is reduced. By plotting the distribution above, we can see where any particular shoe size
will be located on the normal distribution. These distributions also exist for other
measures such as height, weight, and even blood pressure. Similar principles can be
established for psychometric tests and questionnaires which are measuring a particular
ability or trait. Using the information of a distribution we can make assessments and even
predictions about outcomes. For instance, shoe makers will cater for the central tendency
of sizes by making a large number of shoes between sizes seven and nine. Few shoes will be
made smaller and larger than this, because the market is relatively smaller. In the case
of psychometric measurements, we can locate a person's performance on the distribution
curve and it will tell us whether he or she scored very well, average, or low on the
quality relative to the distribution of the sample.
Norms therefore provide the following benefits:
! They provide a real life criterion through which we can establish a standard.
! They describe the quality of an individual's performance by comparing his or her
score with the distribution of people similar to him or her.
! They allow us to compare the performance of an individual on two or more
instruments or measures because they represent standard scales of measurement.
Tests and other instruments vary in difficulty and they seldom consist of the same
number of items. Norms overcome these difficulties by providing a standardised way of
viewing performance. No matter how tests may vary in length or difficulty, we can compare
the person's performance against a common standard using norms. Further, we may want to
compare people to a variety of different groups. So, for example, we can have a range of
norms for a test varying from matriculants, through Technicon graduates, to University
graduates. We can also develop norms for cultural groups where we know the instruments we
are using are influenced by cultural factors. Separate norms may provide an unbiased way
of assessing people on the same instrument and compensating for cultural bias in the
The process utilised in NormMaker smooths and normalises the distribution of a
set of scores to provide a norm table. The norm table can then be used to locate the raw
score, and identify where the person falls in the distribution. The norm tables that are
produced by NormMaker provide three ways in which we can view the distribution. These are:
! Standard Scores
! Percentile Ranks
3.1.1 Standard Scores
Standard scores are normalised raw scores with a mean of 50 and a standard
deviation of 10. In calculating the Standard scores the raw scores are normalised using
Hasting's transformation (see Hasting, 1955, p.192). In addition, the distribution is
smoothed to eliminate spurious scores. These standard scores provide the basis for the
calculation of other representations of distribution such as stanines.
3.1.2 Percentile Rank
The percentile rank associated with a raw score is a popular way of describing a
candidates performance relative to the norm group. The percentile rank tells us that the
raw score in question is higher than or equal to that percentage of scores in the norm
group. For example, a percentile rank of 76 on an assertiveness questionnaire would mean
that the raw score the person obtained shows that he or she is as or more assertive (as
measured by the questionnaire) than 76% of the candidates comprising the norm group.
The most widely used measure of distribution of results for psychometric tests and
questionnaires are stanines. They are derived from standard scores and have a mean of 5
and a standard deviation of 1,96. Stanines are a convenient way of dividing up a normal
distribution into a nine point scale. The stanine scale is particularly useful in that it
eliminates the over interpretation of small spurious differences in scores between
candidates. We can show the way stanines reflect the distribution in the following way:
|In the lowest 0-4% of
||Very low or very poor.
|Better than 4-11% of the
||Low or poor.
|Better than 11-23% of
||Well below average.
|Better than 23-40% of
|In the middle 20% of the
|Better than 60-77% of
|Better than 77-89% of
||High or good
|Better than 89-96% of
||Very high or very good
|In the top 96-100% of